Seminar / Colloquium

Statistics Seminar: How much data is sufficient to learn high-performing algorithms?

Topic: 
How much data is sufficient to learn high-performing algorithms?
Abstract / Description: 

Algorithms often have tunable parameters that have a considerable impact on their runtime and solution quality. A growing body of research has demonstrated that data-driven algorithm design can lead to significant gains in runtime and solution quality. Data-driven algorithm design uses a training set of problem instances sampled from an unknown, application-specific distribution and returns a parameter setting with strong average performance on the training set. We provide a broadly applicable theory for deriving generalization guarantees for data-driven algorithm design, which bound the difference between the algorithm's expected performance and its average performance over the training set.

The challenge is that for many combinatorial algorithms, performance is a volatile function of the parameters: slightly perturbing the parameters can cause a cascade of changes in the algorithm’s behavior. Prior research has proved generalization bounds by employing case-by-case analyses of parameterized greedy algorithms, clustering algorithms, integer programming algorithms, and selling mechanisms. We uncover a unifying structure which we use to prove very general guarantees, yet we recover the bounds from prior research. Our guarantees apply whenever an algorithm's performance is a piecewise-constant, -linear, or — more generally — piecewise-structured function of its parameters. As we demonstrate, our theory also implies novel bounds for dynamic programming algorithms used in computational biology and voting mechanisms from economics.

This talk is based on joint work with Nina Balcan, Dan DeBlasio, Travis Dick, Carl Kingsford, and Tuomas Sandholm.

Date and Time: 
Tuesday, July 20, 2021 - 4:30pm

US-ATMC Fall 2021 Public Seminar Series

Topic: 
“Mobility: Asia Moves Forward in the 4th Industrial Revolution”
Abstract / Description: 

Mobility: Asia Moves Forward in the 4th Industrial Revolution

The shift of IT to the cloud and the growth of new tools such as AI and edge computing have already caused major advances in mobility; for example, smartphones now serve as platforms for banking, education, ride hailing, ID authentication, wellness monitoring, etc. This series examines new technology-and-business solutions that may shape the future of mobility, e.g. smart city infrastructure for autonomous vehicles, intelligent prosthetics for physical mobility, autonomous delivery robots, new propulsion and navigation systems, new applications of mobile IT devices, and more.


The series begins September 23, 2021 and continues weekly until December 2, 2021. Seminars will be held online via Zoom and in-person for students, Thursdays, 5:30 PM - 7:00 PM, with informal networking following until approximately 7:30 PM (PST).

Date and Time: 
Thursday, September 23, 2021 - 5:30pm
Thursday, September 30, 2021 - 5:30pm
Thursday, October 7, 2021 - 5:30pm
Thursday, October 14, 2021 - 5:30pm

STEM to SHTEM final presentations

Topic: 
Final internship presentations
Abstract / Description: 

Join 2021 STEM to SHTEM Interns for their final presentations

STEM to SHTEM program goals are to provide high school and community college students early exposure to cutting-edge research in an academic environment, introduce students to the diversity of research topics pursued in engineering, and emphasize the inseparability of the humanities and the human element from research in STEM.

 

 

Projects from previous years are available via the link below.

Date and Time: 
Thursday, August 12, 2021 - 1:00pm
Venue: 
Zoom ID: 301805583; password required

IT-Forum: "Between tractable and intractable problems in reinforcement learning"

Topic: 
Between tractable and intractable problems in reinforcement learning
Abstract / Description: 

Markov decision processes (MDPs) capture some of the most important aspects of decision making under uncertainty and as such they are at the heart of many efforts to decision making under uncertainty. However, MDPs are "flat" with no structure and as such, planning and learning in MDPs with multidimensional state spaces, common in applications, is provably intractable. Yet, reinforcement learning methods have been quite successful in providing strong solutions to some of these seemingly intractable problems. In this talk I will present my view of how to think about these successes by presenting a framework where the key idea is to give algorithms hints that can create backdoors to crack otherwise intractable problems. The talk will then dive into categorizing hints based on whether they can indeed succeed at doing this for the special case when the hints are given in the form of constraints on how value functions look like in the context of planning with generative models, also known as simulation optimization. As we shall see, seemingly minor differences between hints can cause some hints to work, while others fail.

Date and Time: 
Thursday, June 10, 2021 - 4:30pm

eWEAR Seminar" "Considering Human Perception in User Experience"

Topic: 
Considering Human Perception in User Experience
Abstract / Description: 

9:00 am - 9:30 am - "Understanding the human visual system for more immersive/perceptually realistic virtual and augmented reality displays". Brooke Krajancich, Ph.D. Candidate In Electrical Engineering, Stanford University

9:30 am - 10:00 am - "Prototyping interactions with cars and robots". David Sirkin, Executive Director, Interaction Design, CDR, Stanford University

Date and Time: 
Thursday, June 10, 2021 - 10:00am to Friday, June 11, 2021 - 9:55am
Venue: 
Details will be sent to attendees upon registration

SCIEN Colloquium and EE 292E presents "Learning-based 3D Computer-Generated Holography"

Topic: 
Learning-based 3D Computer-Generated Holography
Abstract / Description: 

Computer generated holography (CGH) is fundamental to applications such as biosensing, volumetric display, optical/acoustic tweezer, security and many others that require spatial control of intricate optical or acoustic fields. For near-eye displays, CGH provides the opportunity to support true 3D projection in a sunglass-like display. Yet, the conventional approach to compute a true 3D hologram via physical simulation of diffraction and inference is slow and unaware of occlusion. Moreover, experimental results are often inferior to simulations due to non-idealized optical systems, non-linear and non-uniform SLM responses, and image degradation caused by complex to phase-only conversion. These computational and hardware-imposed challenges together limit the interactiveness and realism of the ultimate immersive experience. In this talk, I will describe techniques to mitigate these challenges, including physical simulation algorithms that handle occlusion for RGB-D and more advanced 3D input, methods to create large-scale 3D hologram datasets, training of CNNs to speed up complex and phase-only hologram synthesis, and approaches to compensate hardware limitations. Together, the resulted system can synthesis and display photorealistic 3D holograms in real-time using a single consumer-grade GPU and run interactively on an iPhone leveraging the Neural Engine. I will further discuss possible extensions that could be built top of the proposed system to support foveated rendering, static pupil expansion, view-dependent effect and other features.

Date and Time: 
Wednesday, June 2, 2021 - 4:30pm

CANCELLED - SystemX Alliance presents "Autonomous Navigation of Miniature Distributed Space Systems"

Topic: 
Autonomous Navigation of Miniature Distributed Space Systems
Abstract / Description: 

Two key trends are revolutionizing the way humans conduct spaceflight, namely, the miniaturization of satellites (e.g., micro- and nano-satellites) and the distribution of payload tasks among multiple coordinated units (e.g., formation-flying, on-orbit servicing, fractionation, swarms). The combination of these approaches promises breakthroughs in space science (e.g., imaging of earth-like planets, characterization of gravitational waves), remote sensing (e.g., synthetic aperture radar interferometry, aeronomy, gravimetry), and space exploration (e.g., lifetime extension, assembly of structures, space debris removal). Irrespective of the specific application, future miniature distributed space missions require a high level of autonomy to maintain and reconfigure the relative motion of the participating vehicles within the prescribed accuracy and range of operations. Especially on small spacecraft, these requirements are hard to meet due to the limited resources, and the chief goal of current research and development is to pave the way for the autonomous Guidance, Navigation, & Control (GN&C) of "self-driving nanosatellites."

Leveraging the presenter's experience and contributions to recent satellite formation-flying and rendezvous missions in low earth orbit (TanDEM-X, Prisma, Biros), this presentation addresses the navigation algorithms under developments to enable a new class of miniature distributed space instruments (Starling, Visors, Swarm-Ex). The focus will be on autonomous vision-based and radio-frequency spaceborne absolute and relative navigation systems, including their training and validation using high-fidelity hardware-in-the-loop simulations and robotic testbeds at the Stanford's Space Rendezvous Laboratory.

Date and Time: 
Thursday, June 3, 2021 - 4:30pm

Pages

Applied Physics / Physics Colloquium

Applied Physics/Physics Colloquium presents "Fast Radio Bursts: A Cosmological Conundrum or Why I Finally Left the Galaxy"

Topic: 
Fast Radio Bursts: A Cosmological Conundrum or Why I Finally Left the Galaxy
Abstract / Description: 

Fast Radio Bursts (FRBs) are a newly recognized and puzzling phenomenon consisting of few-millisecond bursts of radio waves coming from cosmological distances. They are ubiquitous, with all-sky rates of roughly 1000 per day, signaling a not uncommon yet powerful explosion of presently unknown origin. Recently there has been significant observational progress on understanding FRBs and recognizing their potential as novel cosmological probes, in large part thanks to the Canadian Hydrogen Intensity Mapping Experiment Fast Radio Burst Project (CHIME/FRB). In this talk I review what is known about FRBs, and describe some of the latest CHIME/FRB results related to FRB origin and distribution in the Universe.

Date and Time: 
Tuesday, June 1, 2021 - 4:30pm

Applied Physics/Physics Colloquium presents "Electrons in 2D semiconductor moiré superlattices"

Topic: 
Electrons in 2D semiconductor moiré superlattices
Abstract / Description: 

When two van der Waals materials of slightly different orientations or lattice constants are overlaid, a moiré pattern emerges. The moiré pattern introduces a new length scale, many times the lattice constant of the original materials, for Bragg scattering of Bloch electrons in each layer.  This gives rise to moiré minibands and rich emergent phenomena. In this talk I will discuss recent experiments on angle-aligned semiconductor heterobilayers, which exhibit remarkable correlated insulating states. I will also discuss the prospect of using moiré superlattices as a Hubbard model quantum simulator.

Date and Time: 
Tuesday, May 18, 2021 - 4:30pm

Applied Physics/Physics Colloquium: Imaging the dark matter with millions of stars

Topic: 
Imaging the dark matter with millions of stars
Abstract / Description: 

The ESA Gaia mission has delivered amazingly accurate spatial and kinematic measurements for millions of stars, and rapidly improving information on a billion more. Ground-based spectroscopic surveys deliver additional important information about the birth conditions of these stars. Orbital times in the Galaxy are measured in tens of millions of years, so how do we use this snapshot of kinematic and birth information to learn the gravitational force law in the Milky Way, and, in turn, the distribution of dark matter? There are classical stellar-dynamics approaches that (more-or-less) find consistency relationships among second moments (velocity dispersions). But new approaches have appeared in the last few years that deliver images of stellar orbits. I discuss these new orbit-imaging methods and their promise and limitations. They might end up being the most precise tools we will ever have for constraining the dark matter and its dynamics.

Date and Time: 
Tuesday, May 4, 2021 - 4:30pm

Applied Physics/Physics Colloquium welcomes Xiaoxing Xi

Topic: 
Scientific espionage, open exchange, and American competitiveness
Abstract / Description: 

Amid rapidly escalating tension between the United States and China, professors, scientists, and students of Chinese ethnic origin as well as those engaging in academic collaborations with China are under heightened scrutiny by the federal government. In 2015, Dr. Xi became a casualty of this campaign despite being innocent. This experience gave him insights into the challenges Chinese scientists face and the immediate threat to the open environment in fundamental research.

Date and Time: 
Tuesday, April 27, 2021 - 4:30pm

Applied Physics/Physics Colloquium presents "The NANOGrav 12.5-Year Data Set: Knocking at the Gravitational Wave Door"

Topic: 
The NANOGrav 12.5-Year Data Set: Knocking at the Gravitational Wave Door
Abstract / Description: 

Millisecond pulsars are rapidly rotating neutron stars with phenomenal rotational stability. The NANOGrav collaboration monitors an array of about 80 of these cosmic clocks in order to detect perturbations due to gravitational waves at nanohertz frequencies. These gravitational waves will most likely result from an ensemble of supermassive black hole binaries. Their detection and subsequent study will offer unique insights into galaxy growth and evolution over cosmic time. I will present our most recent dataset and the results of our gravitational wave analysis, which suggests the presence of a common spectral signature in the data that could be the first hints of a gravitational wave background. I will then describe the gains in sensitivity that are expected from additional data, discoveries of millisecond pulsars, more sensitive instrumentation, and international collaboration and discuss prospects for detection in the next several years.

Date and Time: 
Tuesday, April 20, 2021 - 4:30pm

Applied Physics/Physics Colloquium presents "Spin dynamics of ultracold atoms in optical lattices"

Topic: 
Spin dynamics of ultracold atoms in optical lattices
Abstract / Description: 

Ultracold atoms offer a unique platform to study spin physics. When atoms are arranged in an optical lattice in form of a Mott insulator, the atomic motion is frozen out and the study and control of the spin degree of freedom emerges as a new frontier. Heisenberg spin models, where only neighboring spins interact, are the paradigmatic model for many interesting phenomena. Until very recently, all experimental studies with cold atoms addressed the special case of an isotropic Heisenberg model. Using lithium-7 atoms and Feshbach resonances to tune the interactions, we have created spin ½ Heisenberg models with adjustable anisotropy, including the special XX-model which can be exactly solved by mapping it to non-interacting fermions. Spin transport changes from ballistic to diffusive depending on the anisotropy. For transverse spin patterns, we have found several new dephasing mechanisms related to a superexchange induced effective magnetic field. Using rubidium atoms and two atoms per site, we have realized spin 1 models. The onsite interactions give rise to a so-called-single-ion anisotropy term proportional to (S_z)^2, which plays an important role in stabilizing magnetism for low-dimensional magnetic materials. Our studies of spin dynamics illustrate the role of ultracold atoms as a quantum simulator for materials and for elucidating fundamental aspects of many-body physics.

 

Date and Time: 
Tuesday, April 13, 2021 - 4:30pm

2021 Robert Hofstadter Memorial Lecture: What happened to the kilogram?

Topic: 
What happened to the kilogram?
Abstract / Description: 

For 130 years, a cylinder made of a platinum-iridium alloy stored near Paris was the official definition of a kilogram, the basic unit of mass. This all changed on May 20, 2019: a kilogram is now defined by a fundamental constant of nature known as the Planck constant h, which relates the energy of a photon to its frequency: h= 6.62607015 10-34 kilograms times square meters per second.
Sounds complicated? In this talk, I will provide the reasons for changing the definition of the kilogram, give simple explanations what the new kilogram is conceptually, and explain how objects with exactly known masses can be realized using advanced technology.


We are excited to announce that our distinguished speaker is Prof. Wolfgang Ketterle, the John D. MacArthur Professor of Physics at the Massachusetts Institute of Technology. Prof. Ketterle works on experimental research in atomic physics and laser spectroscopy and focuses currently on Bose-Einstein condensation in dilute atomic gases. He was among the first scientists to observe this phenomenon in 1995, and realized the first atom laser in 1997. His earlier research was in molecular spectroscopy and combustion diagnostics.

Prof. Ketterle's awards include the Rabi Prize of the American Physical Society (1997), the Gustav-Hertz Prize of the German Physical Society (1997), the Discover Magazine Award for Technological Innovation (1998), the Fritz London Prize in Low Temperature Physics (1999), the Dannie-Heineman Prize of the Academy of Sciences, Göttingen, Germany (1999), the Benjamin Franklin Medal in Physics (2000), and the Nobel Prize in Physics (2001, together with E.A. Cornell and C.E. Wieman).

Prof. Ketterle will also give the Applied Physics/Physics colloquium at 4:30 PM on Tuesday, April 13, 2021. Detailed information about both lectures and our speaker can be found on Physics site (link below). As in previous years, both lectures are free and open to the public via Zoom webinar.

Date and Time: 
Monday, April 12, 2021 - 4:30pm
Venue: 
Zoom ID: 95238055074; Passcode: 642168

Applied Physics/Physics Colloquium presents "Black holes, information and wormholes"

Topic: 
Black holes, information and wormholes
Abstract / Description: 

In the 1970s, Hawking showed that black holes, like any finite-temperature system, radiate energy and so eventually evaporate away entirely. However, his calculations showed something very weird: unlike any other physical system, the radiation seemed to contain no information about the initial state of the black hole. Instead the information that fell into the black hole was simply lost forever. This contradiction between Hawking's calculations and the ordinary rules of quantum mechanics has been a driving force behind much of the research in quantum gravity over the ensuing decades. Finally, in the last couple of years, we have begun to understand where Hawking's calculation went wrong, and to derive precise predictions, consistent with unitary quantum mechanics, for the information content of Hawking radiation. However, the new calculations, which involve weird spacetime topologies called 'spacetime wormholes', lead to as many new questions as answers.

Date and Time: 
Tuesday, April 6, 2021 - 4:30pm

Pages

CS300 Seminar

#StanfordToo: A Conversation about Sexual Harassment in Our Academic Spaces

Topic: 
#StanfordToo: A Conversation about Sexual Harassment in Our Academic Spaces
Abstract / Description: 

Individuals of all genders invited to be a part of:
#StanfordToo: A Conversation about Sexual Harassment in Our Academic Spaces, where we will feature real stories of harassment at Stanford academic STEM in a conversation with Provost Drell, Dean Minor (SoM), and Dean Graham (SE3). We will have plenty of time for audience discussion on how we can take concrete action to dismantle this culture and actively work towards a more inclusive Stanford for everyone. While our emphasis is on STEM fields, we welcome and encourage participation from students, postdocs, staff, and faculty of all academic disciplines and backgrounds.

Date and Time: 
Friday, April 19, 2019 - 3:30pm
Venue: 
STLC 111

John G. Linvill Distinguished Seminar on Electronic Systems Technology

Topic: 
Internet of Things and Internet of Energy for Connecting at Any Time and Any Place
Abstract / Description: 

In this presentation, I would like to discuss with you how to establish a sustainable and smart society through the internet of energy for connecting at any time and any place. I suspect that you have heard the phrase, "Internet of Energy" less often. The meaning of this phrase is simple. Because of a ubiquitous energy transmission system, you do not need to worry about a shortage of electric power. One of the most important items for establishing a sustainable society is [...]


"Inaugural Linvill Distinguished Seminar on Electronic Systems Technology," EE News, July 2018

 

Date and Time: 
Monday, January 14, 2019 - 4:30pm
Venue: 
Hewlett 200

Special Seminar: Formal Methods meets Machine Learning: Explorations in Cyber-Physical Systems Design

Topic: 
Formal Methods meets Machine Learning: Explorations in Cyber-Physical Systems Design
Abstract / Description: 

Cyber-physical systems (CPS) are computational systems tightly integrated with physical processes. Examples include modern automobiles, fly-by-wire aircraft, software-controlled medical devices, robots, and many more. In recent times, these systems have exploded in complexity due to the growing amount of software and networking integrated into physical environments via real-time control loops, as well as the growing use of machine learning and artificial intelligence (AI) techniques. At the same time, these systems must be designed with strong verifiable guarantees.

In this talk, I will describe our research explorations at the intersection of machine learning and formal methods that address some of the challenges in CPS design. First, I will describe how machine learning techniques can be blended with formal methods to address challenges in specification, design, and verification of industrial CPS. In particular, I will discuss the use of formal inductive synthesis --- algorithmic synthesis from examples with formal guarantees — for CPS design. Next, I will discuss how formal methods can be used to improve the level of assurance in systems that rely heavily on machine learning, such as autonomous vehicles using deep learning for perception. Both theory and industrial case studies will be discussed, with a special focus on the automotive domain. I will conclude with a brief discussion of the major remaining challenges posed by the use of machine learning and AI in CPS.

Date and Time: 
Monday, December 4, 2017 - 4:00pm to Tuesday, December 5, 2017 - 3:55pm
Venue: 
Gates 463A

SpaceX's journey on the road to mars

Topic: 
SpaceX's journey on the road to mars
Abstract / Description: 

SSI will be hosting Gwynne Shotwell — President and COO of SpaceX — to discuss SpaceX's journey on the road to mars. The event will be on Wednesday Oct 11th from 7pm - 8pm in Dinkelspiel Auditorium. After the talk, there will be a Q&A session hosted by Steve Jurvetson from DFJ Venture Capital.

Claim your tickets now on eventbright

 

Date and Time: 
Wednesday, October 11, 2017 - 7:00pm to Thursday, October 12, 2017 - 6:55pm
Venue: 
Dinkelspiel Auditorium

CS Department Lecture Series (CS300)

Topic: 
Faculty speak about their research to new PhD students
Abstract / Description: 

Offered to incoming first-year PhD students in the Autumn quarter.

The seminar gives CS faculty the opportunity to speak about their research, which allows new CS PhD students the chance to learn about the professors and their research before permanently aligning.

4:30-5:15, Subhasish Mitra

5:15-6:00, Silvio Savarese

Date and Time: 
Wednesday, December 7, 2016 - 4:30pm to Thursday, December 8, 2016 - 5:55pm
Venue: 
200-305 Lane History Corner, Main Quad

CS Department Lecture Series (CS300)

Topic: 
Faculty speak about their research to new PhD students
Abstract / Description: 

Offered to incoming first-year PhD students in the Autumn quarter.

The seminar gives CS faculty the opportunity to speak about their research, which allows new CS PhD students the chance to learn about the professors and their research before permanently aligning.

4:30-5:15, Phil Levis

5:15-6:00, Ron Fedkiw

Date and Time: 
Monday, December 5, 2016 - 4:30pm to Tuesday, December 6, 2016 - 5:55pm
Venue: 
200-305 Lane History Corner, Main Quad

CS Department Lecture Series (CS300)

Topic: 
Faculty speak about their research to new PhD students
Abstract / Description: 

Offered to incoming first-year PhD students in the Autumn quarter.

The seminar gives CS faculty the opportunity to speak about their research, which allows new CS PhD students the chance to learn about the professors and their research before permanently aligning.

4:30-5:15, Dan Boneh

5:15-6:00, Aaron Sidford

Date and Time: 
Wednesday, November 30, 2016 - 4:30pm to Thursday, December 1, 2016 - 5:55pm
Venue: 
200-305 Lane History Corner, Main Quad

CS Department Lecture Series (CS300)

Topic: 
Faculty speak about their research to new PhD students
Abstract / Description: 

Offered to incoming first-year PhD students in the Autumn quarter.

The seminar gives CS faculty the opportunity to speak about their research, which allows new CS PhD students the chance to learn about the professors and their research before permanently aligning.

4:30-5:15, John Mitchell

5:15-6:00, James Zou

Date and Time: 
Monday, November 28, 2016 - 4:30pm to Tuesday, November 29, 2016 - 5:55pm
Venue: 
200-305 Lane History Corner, Main Quad

CS Department Lecture Series (CS300)

Topic: 
Faculty speak about their research to new PhD students
Abstract / Description: 

Offered to incoming first-year PhD students in the Autumn quarter.

The seminar gives CS faculty the opportunity to speak about their research, which allows new CS PhD students the chance to learn about the professors and their research before permanently aligning.

4:30-5:15, Emma Brunskill

5:15-6:00, Doug James

Date and Time: 
Wednesday, November 16, 2016 - 4:30pm to Thursday, November 17, 2016 - 5:55pm
Venue: 
200-305 Lane History Corner, Main Quad

CS Department Lecture Series (CS300)

Topic: 
Faculty speak about their research to new PhD students
Abstract / Description: 

Offered to incoming first-year PhD students in the Autumn quarter.

The seminar gives CS faculty the opportunity to speak about their research, which allows new CS PhD students the chance to learn about the professors and their research before permanently aligning.

4:30-5:15, James Landay

5:15-6:00, Dan Jurafsky

Date and Time: 
Monday, November 14, 2016 - 4:30pm to Tuesday, November 15, 2016 - 5:55pm
Venue: 
200-305 Lane History Corner, Main Quad

Pages

EE380 Computer Systems Colloquium

SystemX presents "Quantum Computing – Where does the Hope meet the Hype?"

Topic: 
Quantum Computing – Where does the Hope meet the Hype?
Abstract / Description: 

Quantum computers exploit the bizarre features of quantum physics -- uncertainty, entanglement, and measurement -- to perform tasks that are impossible using conventional means, such as computing over ungodly amounts of data, and communicating via teleportation. I will describe the architecture of a quantum computer based on individual atomic clock qubits, suspended and isolated with electric fields, perfectly replicable with no idle errors, and individually addressed with laser beams. This leading physical representation of a quantum computer has allowed unmatched demonstrations of small algorithms and emulations of hard quantum problems with more than 50 quantum bits. While this system can solve some esoteric tasks that cannot be accomplished in conventional devices, it remains a great engineering challenge to build a quantum computer big enough to be generally useful for society. But the good news is that this is not a scientific challenge, as we know the technology needed and it's not quantum.

Date and Time: 
Thursday, October 8, 2020 - 4:30pm
Venue: 
Zoom ID: 920 2334 9868; +passcode

EE380 Computer Systems Colloquium presents "Coming Attractions: Death or Utopia in the Next Three Decades"

Topic: 
Coming Attractions: Death or Utopia in the Next Three Decades
Abstract / Description: 

Today the data suggests that we are near the beginning of a chaotic mess of global proportions. Things are fairly simple: a global pandemic with no tools to fight the virus. a global economy in disarray, climate change and other existential risks beginning to intrude into our daily lives, and a total lack of a plan as to what to do. On the other hand, we are at the pinnacle of human capabilities and have, if we choose to do so, the capability create a Utopian egalitarian world without conflict or want.

In this two hour program a group of experts will explore the future, focusing on 2030 and 2050. Where are we now? What is trending? What if anything can be done about it?


Sponsors: The invitational Asilomar Microcomputer Workshop is one of the iconic gatherings which supported the growth of computing. This is the first mini-conference which replaces the 46th Asilomar Microcomputer Workshop which was canceled due to the COVID-19 pandemic. www.amw.org.

Date and Time: 
Thursday, June 4, 2020 - 11:00am to Friday, June 5, 2020 - 10:55am
Venue: 
REGISTER to receive access

EE380 Computer Systems Colloquium presents "Rebooting the Internet"

Topic: 
Rebooting the Internet
Abstract / Description: 

"Build one and throw it out," so the adage goes. This talk explores the argument for systems architecture revision, including processors, operating systems, and networking--both as a general principle and the ways the Internet in particular is currently in need of a reboot. Assumptions, resources, and goals change with time and experience, and so too does our understanding of architectural principles. Through the eyes of the Internet and other examples, we review what we got right (one ring to rule them all, good enough rather than perfect), what we got wrong (7 layers, name resolution as afterthought), and what we only now are beginning to appreciate (layering and forwarding as one thing). Challenge cases are presented that can help drive this redesign, including single-packet exchanges and recursive layering, and an example given of one direction this approach can lead. Finally, we explore the challenges of evolution and transition to help us prepare for giving the Internet a well-deserved reset.

REMOTE URL SU-EE380-20200311

Date and Time: 
Wednesday, March 11, 2020 - 4:30pm
Venue: 
[remote only]

EE380 Computer Systems Colloquium presents "Data Analytics at the Exascale for Free Electron Lasers Lasers Project"

Topic: 
Data Analytics at the Exascale for Free Electron Lasers Lasers Project
Abstract / Description: 

The increase in velocity, volume, and complexity of the data generated by the upcoming Linac Coherent Light Source upgrade (LCLS-II) at the SLAC National Accelerator Laboratory presents a considerable challenge for data acquisition, data processing, and data management. These systems face formidable challenges due to the extremely high data throughput, hundreds of GB/s to multi-TB/s, generated by the detectors at the experimental facilities and to the intensive computational demand for data processing and scientific interpretation. The LCLS-II Data System is a fast, powerful, and flexible architecture that includes a feature extraction layer designed to reduce the data volumes by at least one order of magnitude while preserving the science content of the data. Innovative architectures are required to implement this reduction with a configurable approach that can adapt to the multiple science areas served by LCLS. In order to increase the likelihood of experiment success and improve the quality of recorded data, a real-time analysis framework provides visualization and graphically-configurable analysis of a selectable subset of the data on the timescale of seconds. A fast feedback layer offers dedicated processing resources to the running experiment in order to provide experimenters feedback about the quality of acquired data within minutes. We will present an overview of the LCLS-II Data System architecture with an emphasis on the Data Reduction Pipeline (DRP) and online monitoring framework.

Date and Time: 
Wednesday, March 4, 2020 - 4:30pm
Venue: 
Shriram 104

EE380 Computer Systems Colloquium presents "The Soul of a New Machine: Rethinking the Computer"

Topic: 
The Soul of a New Machine: Rethinking the Computer
Abstract / Description: 

While our software systems have become increasingly elastic, the physical substrate available to run that software (that is, the computer!) has remained stuck in a bygone era of PC architecture. Hyperscale infrastructure providers have long since figured this out, building machines that are fit to purpose -- but those advances have been denied to the mass market. In this talk, we will talk about our vision for a new, rack-scale, server-side machine -- and how we anticipate advances like open firmware, RISC-V, and Rust will play a central role in realizing that vision.

Date and Time: 
Wednesday, February 26, 2020 - 4:30pm
Venue: 
Shriram 104

EE380 Computer Systems Colloquium presents "fastai: A Layered API for Deep Learning"

Topic: 
fastai: A Layered API for Deep Learning
Abstract / Description: 


[speaker photo] Sylvain's research at fast.ai has focused on designing and improving techniques that allow models to train fast with limited resources. He has also been a core developer of the fastai library, including implementing the warping transformations, the preprocessing pipeline, much of fastai.text, and a lot more.
Prior to fastai, Sylvain was a Mathematics and Computer Science teacher in Paris for seven years. He taught CPGE, the 2-year French program that prepares students for graduate programs at France's top engineering schools (the "grandes écoles"). After relocating to the USA in 2015, Sylvain wrote ten textbooks covering the entire CPGE curriculum. Sylvain is an alumni from École Normale Supérieure (Paris, France) and has a Master's Degree in Mathematics from University Paris XI (Orsay, France). He lives in Brooklyn with his husband and two sons.fastai is a deep learning library which provides practitioners with high-level components that can quickly and easily provide state-of-the-art results in standard deep learning domains, and provides researchers with low-level components that can be mixed and matched to build new approaches. It aims to do both things without substantial compromises in ease of use, flexibility, or performance. This is possible thanks to a carefully layered architecture, which expresses common underlying patterns of many deep learning and data processing techniques in terms of decoupled abstractions. These abstractions can be expressed concisely and clearly by leveraging the dynamism of the underlying Python language and the flexibility of the PyTorch library. fastai includes: a new type dispatch system for Python along with a semantic type hierarchy for tensors; a GPU-optimized computer vision library which can be extended in pure Python; an optimizer which refactors out the common functionality of modern optimizers into two basic pieces, allowing optimization algorithms to be implemented in 4-5 lines of code; a novel 2-way callback system that can access any part of the data, model, or optimizer and change it at any point during training; a new data block API; and much more. We have used this library to successfully create a complete deep learning course, which we were able to write more quickly than using previous approaches, and the code was more clear. The library is already in wide use in research, industry, and teaching.

Date and Time: 
Wednesday, February 19, 2020 - 4:30pm
Venue: 
Shriram 104

EE380 Computer Systems Colloquium presents "Centaur Technology's Deep learning Coprocessor"

Topic: 
Centaur Technology's Deep learning Coprocessor
Abstract / Description: 

This talk will explore Centaur's microprocessor-design technology includes both a new high-performance x86 core AND the industry's FIRST integrated AI Coprocessor for x86 systems. The x86 microprocessor cores deliver high instructions/clock (IPC) for server-class applications and support the latest x86 extensions such as AVX 512 and new instructions for fast transfer of AI data. The AI Coprocessor is a clean-sheet processor designed to deliver high performance and efficiency on deep-learning applications, freeing up the x86 cores for general-purpose computing. Both the x86 and AI Coprocessor technologies have now been proven in silicon using a new scalable SoC platform with eight x86 cores and an AI Coprocessor able to compute 20 trillion AI operations/sec with 20 terabytes/sec memory bandwidth. This SoC architecture requires less than 195mm2 in TSMC 16nm and provides an extensible platform with 44 PCIe lanes and 4 channels of PC3200 DDR4.

Date and Time: 
Wednesday, February 12, 2020 - 4:30pm
Venue: 
Shriram 104

EE380 Computer Systems Colloquium presents "KUtrace 2020"

Topic: 
KUtrace 2020
Abstract / Description: 

Observation tools for understanding occasionally-slow performance in large-scale distributed transaction systems are not keeping up with the complexity of the environment. The same applies to large database systems, to real-time control systems in cars and airplanes, and to operating system design.

Extremely low-overhead tracing can reveal the true execution and non-execution (waiting) dynamics of such software, running in situ with live traffic. KUtrace is such a tool, based on small Linux kernel patches recording and timestamping every transition between kernel- and user-mode execution across all CPUs of a datacenter or vehicle computer. The resulting displays show exactly what each transaction is doing every nanosecond, and hence shows why unpredictable ones are slow, all with tracing overhead well under 1%. Recent additions to KUtrace also show interference between programs and show profiles within long execution stretches that have no transitions.

The net result is deep insight into the dynamics of complex software, leading to often-simple changes to improve performance.

Date and Time: 
Wednesday, February 5, 2020 - 4:30pm
Venue: 
Shriram 104

EE380 Computer Systems Colloquium presents "Computer Security: The Mess We're In, How We Got Here, and What to Do About It"

Topic: 
Computer Security: The Mess We're In, How We Got Here, and What to Do About It
Abstract / Description: 

There are dozens of companies selling products to solve your Identity and Access Management (IAM) problems. Why so many? Because IAM is really hard. But why is IAM so hard? I hope to convince you in this talk the reason is that our approach to IAM is fundamentally flawed. The talk includes my personal theory of how we got into this mess and a strategy for getting out of it.

Date and Time: 
Wednesday, January 29, 2020 - 4:30pm
Venue: 
Shriram 104

EE380 Computer Systems Colloquium presents "A Fight over the Law of Software APIs, and other war stories from the Electronic Frontier Foundation""

Topic: 
A Fight over the Law of Software APIs, and other war stories from the Electronic Frontier Foundation
Abstract / Description: 

Over the past four years, a federal appeals court has upended a basic assumption about software: that while software can be covered by copyright law, software interfaces cannot. The court ruled that Google's re-implementation of Java APIs infringed Oracle's Java copyrights. Now, the Supreme Court will weigh in. The Electronic Frontier Foundation has argued at each stage of the case that copyright in APIs threatens innovation, especially by smaller players. This is part of EFF's work promoting the freedom to tinker, innovate, and research, especially outside the walls of major corporations—work that has also included challenges to overbroad computer crime and "anti-circumvention" laws.

Date and Time: 
Wednesday, January 22, 2020 - 4:30pm
Venue: 
Shriram 104

Pages

Ginzton Lab

SPRC & SOS Seminar: Trace Gas Flux Measurements in Artic Ecosystems – Current Applications and Future Perspectives

Topic: 
Trace Gas Flux Measurements in Artic Ecosystems – Current Applications and Future Perspectives
Abstract / Description: 

The exchanges of carbon dioxide (CO2) and methane (CH4) between arctic ecosystems and the atmosphere are sizeable and may if they change may alter the further development of climate warming. These fluxes take place between the atmospheric background of these trace gases and a very large reservoir of organic carbon stored in arctic soils and sediments. The stored carbon amounts to more than twice the current global atmospheric burden of CO2. Therefore, accurate measurements of the fluxes and how they vary temporally and spatially is key to understanding possible ecosystem impacts on the atmospheric mixing ratios and hence climate.
This presentation will detail the current state of the global CO2 and CH4 budget and review methods applied in quantifying the arctic trace gas flux components. The presentation will show new developments in the attempts to use UAV's to help improve spatial coverage of measured fluxes. The challenges in applying the laser techniques needed for the precise and fast trace gas measurements under true arctic conditions will be discussed.

Date and Time: 
Tuesday, May 4, 2021 - 11:00am to Wednesday, May 5, 2021 - 10:55am

OSA/SPIE present "Observing Ocean Health With A Global Network of Chemical and Biological Sensors on Robotic Platforms: Biogeochemical-Argo"

Topic: 
Observing Ocean Health With A Global Network of Chemical and Biological Sensors on Robotic Platforms: Biogeochemical-Argo
Abstract / Description: 

The ocean is under threat from a variety of processes driven by increasing atmospheric carbon dioxide such as ocean warming, changing winds and currents, decreasing pH and oxygen, and less ice cover at high latitudes. However, vast areas of the open ocean are sampled from research ships for chemical and biological properties only once per decade or less, with sampling occurring mainly in summer. Our ability to detect changes in ocean chemical and biological processes that may be occurring are greatly hindered by this undersampling. Robotic platforms carrying chemical and biological sensors that can monitor ocean metabolism, such as rates of oxygen production, nutrient uptake, accumulation of plankton biomass, and respiration, are required. The platforms and sensors must operate for the 5 to 10 year period between research vessel visits with no direct human intervention and little or no chance for sensor recalibration. Here, I'll describe the BGC-Argo network, which is being implemented to meet this mission. The BGC-Argo array is based on profiling floats that cycle through the upper 2 kilometers of the ocean every 10 days for 4 to 6 years. These floats carry optical sensors for oxygen, nitrate, chlorophyll, and particles, as well as electrochemical sensors for pH and physical sensors for temperature, pressure and salinity. The nitrate and pH sensors used in the array were developed at MBARI. With funding from NSF, NOAA, and NASA, we have deployed 200 floats in the Southern Ocean over the past 6 years. A recent research infrastructure grant from the NSF will extend the Southern Ocean array to the globe with 500 floats additional floats.

Date and Time: 
Tuesday, January 26, 2021 - 1:30pm

OSA/SPIE, SPRC and Ginzton Lab present "Next Generation Photonics"

Topic: 
Next Generation Photonics
Abstract / Description: 

Over the past few decades, silicon photonics has revolutionized photonic integrated circuits by leveraging the semiconductor CMOS manufacturing infrastructure for low cost, high performance devices and systems. However, key fundamental challenges of the field remain unsolved: packaging the devices with optical fibers and generating light on chip.

We developed a novel approach for fiber packaging based on fusing the fiber and chip together. Connecting a silicon photonic chip across long distances requires attaching optical fibers to the chip. In practice, the packaging of optical fibers to photonic devices is time consuming, lossy, and expensive. This process is usually done by gluing the fiber and chip together using optical adhesives. By fusion splicing the chip and fiber together we have demonstrated losses as low as 1dB for single fibers and 2.5dB for an array of four fibers.

Imagine a laser as thick as an atom that is compatible with silicon photonics. Silicon based materials are passive. Since silicon is an indirect band gap material it is a poor light emitter. To generate light on a silicon photonic integrated circuit, you need to integrate active materials, which are usually not compatible with CMOS, with the device. Two-dimensional materials are excellent candidates for light sources, modulators, and detectors; and they are compatible with CMOS electronic manufacturing in the back end. We recently demonstrated the first fully on-chip 2D laser. Due to their thickness and transfer process, we envision electronic-photonic devices with many optical layers, each with their own lasers, modulators, and detectors based on 2D materials. Our recent demonstration completes the set of active devices completely based on 2D materials.


Password: 017994

Date and Time: 
Wednesday, May 6, 2020 - 1:00pm to Thursday, May 7, 2020 - 12:55pm
Venue: 
Zoom ID: 991 2584 9377

OSA/SPIE with SPRC and Ginzton Lab present "Next Generation Photonics"

Topic: 
Next Generation Photonics
Abstract / Description: 

Over the past few decades, silicon photonics has revolutionized photonic integrated circuits by leveraging the semiconductor CMOS manufacturing infrastructure for low cost, high performance devices and systems. However, key fundamental challenges of the field remain unsolved: packaging the devices with optical fibers and generating light on chip.

We developed a novel approach for fiber packaging based on fusing the fiber and chip together. Connecting a silicon photonic chip across long distances requires attaching optical fibers to the chip. In practice, the packaging of optical fibers to photonic devices is time consuming, lossy, and expensive. This process is usually done by gluing the fiber and chip together using optical adhesives. By fusion splicing the chip and fiber together we have demonstrated losses as low as 1dB for single fibers and 2.5dB for an array of four fibers.

Imagine a laser as thick as an atom that is compatible with silicon photonics. Silicon based materials are passive. Since silicon is an indirect band gap material it is a poor light emitter. To generate light on a silicon photonic integrated circuit, you need to integrate active materials, which are usually not compatible with CMOS, with the device. Two-dimensional materials are excellent candidates for light sources, modulators, and detectors; and they are compatible with CMOS electronic manufacturing in the back end. We recently demonstrated the first fully on-chip 2D laser. Due to their thickness and transfer process, we envision electronic-photonic devices with many optical layers, each with their own lasers, modulators, and detectors based on 2D materials. Our recent demonstration completes the set of active devices completely based on 2D materials.

Date and Time: 
Monday, May 4, 2020 - 12:05pm
Venue: 
Zoom ID: 991 2584 9377 (contact organizers for password)

OSA/SPIE, SPRC and Ginzton Lab present "Applications of Precision Motion Control in the Optics and Photonics Industry"

Topic: 
Applications of Precision Motion Control in the Optics and Photonics Industry
Abstract / Description: 

Precision Automation has become a rapidly growing industry as manufacturers must develop ways to achieve ever-tightening tolerances. Likewise, research must push the envelope in order to maintain the increasingly advancing pace of technology. Aerotech has been providing assistance and solutions to researchers within academia and laboratories worldwide for 50 years and continues to empower students and technicians to complete and publish their research. In this seminar, we will discuss why automation and motion control is important as well as the various methods by which Aerotech has helped expand the boundaries of research in the field of optics and photonics. The topics we will cover in this session include:

  • What Precision Motion Control is and why it is important
  • How to choose the right set of motion control
  • How Aerotech has helped further research in:
    • Optics design and manufacturing
    • Light sources and beamlines
    • Optical positioning and ultra-fast spectroscopy
    • Fiber alignment
    • Laser Processing
    • Additive Manufacturing

After the presentation, a round-table discussion will be held during which we'd like to hear from you! Please bring with you any topics and - more importantly - any tough issues that you'd like to pose pertaining to automated motion control within your specific research field.

Date and Time: 
Tuesday, March 3, 2020 - 4:15pm
Venue: 
Spilker 232

OSA/SPIE present "Applications of Precision Motion Control in the Optics and Photonics Industry"

Topic: 
Applications of Precision Motion Control in the Optics and Photonics Industry
Abstract / Description: 

Refreshments at 4:00

Precision Automation has become a rapidly growing industry as manufacturers must develop ways to achieve ever-tightening tolerances. Likewise, research must push the envelope in order to maintain the increasingly advancing pace of technology. Aerotech has been providing assistance and solutions to researchers within academia and laboratories worldwide for 50 years and continues to empower students and technicians to complete and publish their research. In this seminar, we will discuss why automation and motion control is important as well as the various methods by which Aerotech has helped expand the boundaries of research in the field of optics and photonics. The topics we will cover in this session include:

  • What Precision Motion Control is and why it is important
  • How to choose the right set of motion control
  • How Aerotech has helped further research in:
    • Optics design and manufacturing
    • Light sources and beamlines
    • Optical positioning and ultra-fast spectroscopy
    • Fiber alignment
    • Laser Processing
    • Additive Manufacturing

After the presentation, a round-table discussion will be held during which we'd like to hear from you! Please bring with you any topics and - more importantly - any tough issues that you'd like to pose pertaining to automated motion control within your specific research field.

 

Date and Time: 
Wednesday, February 26, 2020 - 4:15pm
Venue: 
Spilker 232

OSA/SPIE, SPRC and Ginzton Lab present "Innovation in the Laser Industry"

Topic: 
Innovation in the Laser Industry
Abstract / Description: 
Spectra-Physics was a startup when Prof. Byer was a summer student in 1964.  By the time Alan Petersen and Jim Kafka joined in the early 1980s, the company had matured and aspired to be all things in lasers.  We will discuss the changes in the laser industry since that time and the role of innovation within the company to drive growth and broaden the product offerings from scientific lasers to industrial laser systems.
 
PLEASE RVP TO GUARANTEE A SPOT.
Date and Time: 
Monday, March 9, 2020 - 12:30pm
Venue: 
Y2E2 299

Q-Farm Quantum Seminar Series presents "Surprises from Time Crystals"

Topic: 
Surprises from Time Crystals
Abstract / Description: 

Time crystals are new states of matter that only exist in an out-of-equilibrium setting. I will review the state of this rapidly evolving field, focusing in particular on some of the remarkable properties of this phase, and the surprises coming out of its study. I will provide a detailed overview of existing experiments, with a view towards identifying the ingredients needed for an unambiguous observation of this phase in the future.

Date and Time: 
Wednesday, January 29, 2020 - 12:00pm to Thursday, January 30, 2020 - 11:55am
Venue: 
Hansen Physics & Astrophysics Building, 102/103

AP 483 Seminar Series presents "Photovoltaic Restoration of Sight in Retinal Degeneration"

Topic: 
Photovoltaic Restoration of Sight in Retinal Degeneration
Abstract / Description: 

Retinal degenerative diseases lead to blindness due to loss of the “image capturing” photoreceptors, while neurons in the “image-processing” inner retinal layers are relatively well-preserved. Information can be reintroduced into the visual system by photovoltaic subretinal implants, which convert incident light into electric current and stimulate the secondary retinal neurons.

 

To provide sufficient light intensity for photovoltaic stimulation while avoiding visual perception by remaining photoreceptors, images captured by a camera are projected onto the retina from augmented-reality glasses using pulsed near-infrared light. This design avoids the use of bulky electronics and wiring, thereby greatly reducing the surgical complexity and enabling scaling the number of photovoltaic pixels to thousands. Many features of the natural retinal signal processing are preserved in this approach, and spatial resolution matches the pixel pitch (so far 100 μm pixels in human patients, and 50 μm in rodents). For a broad acceptance of this technology by patients who lost central vision due to Age-Related Macular Degeneration, visual acuity should exceed 20/100, which requires pixels smaller than 25 μm. I will present a 3-dimensional electro-neural interface scalable to cellular dimensions and discuss the outlook and challenges for future developments.


 AP483 Optics and Electronics Seminar Series 2019-20 (Sponsored by Ginzton Laboratory, SPRC, Applied Physics, Physics, and HEPL).

Date and Time: 
Monday, February 24, 2020 - 4:15pm
Venue: 
Spilker 232

AP 483 Seminar Series presents "Taking the Humble FBG on a Voyage of Discovery from the Lab Bench to the Hospital and Beyond"

Topic: 
Taking the Humble FBG on a Voyage of Discovery from the Lab Bench to the Hospital and Beyond
Abstract / Description: 

Fiber Bragg gratings (FBGs) emerged almost magically in 1978. Since then, they have developed from being primarily of academic interest, to being one of the most versatile photonic components for both telecommunications and sensing. 

I was lucky enough to get introduced to photonics during the early years of discovery and experimentation with FBGs and ended up making FBG components for telecommunications. Jumping forward a few decades, I had moved away from telecommunications and had started to dabble with free space optical sensing. Then, during a chance meeting over a beer in Sydney, I was asked, rather naively, if it was possible to use optical fiber to monitor what goes on in the esophagus when we swallow. This set me off on an entirely new path and, pulling together some ideas from telecoms, and some basic mechanical engineering, my team ended up developing a range of sensors for monitoring pressure in the human digestive tract. 

Being able to detect swallowing disorders quickly led to monitoring in other regions of the gut, like the colon and small bowel, and, together with colleagues from Flinders Medical Center, we provided some details of the inner workings of the human gastrointestinal tract. Thus, the fiber optic catheter was born. This device kept me busy for almost 10 years, during which time we worked closely with clinical research groups to write a whole new chapter on how the gut works.

The next ‘Eureka’ moment came when we had to develop a temperature independent version of our sensor, initially for monitoring pressure beneath bandages. The very simple design that resulted worked well for sub-bandage measurements but has also become a key technology that has moved to applications in aerospace, pipeline monitoring, and mining.

When I started working in optics, I never thought I'd end up monitoring what makes your stomach growl when hungry, how air flows across an airplane wing, or detecting pressure transients in water pipes. 

During this talk I will explain how our basic transducers work and will then describe the applications they are now being applied to, demonstrating how the humble FBG has opened up the scope and reach of fiber-optic sensing.


 

AP483 Optics and Electronics Seminar Series 2019-20 (Sponsored by Ginzton Laboratory, SPRC, Applied Physics, Physics, and HEPL).

Date and Time: 
Monday, February 10, 2020 - 4:15pm
Venue: 
Spilker 232

Pages

Information Systems Lab (ISL) Colloquium

ISL Colloquium presents "The Hidden Convex Optimization Landscape of Deep Neural Networks"

Topic: 
The Hidden Convex Optimization Landscape of Deep Neural Networks
Abstract / Description: 

The popularity of Deep Neural Networks (DNNs) continues to grow as a result of the great empirical success in a large number of machine learning tasks. However, despite their prevalence in machine learning and the dramatic surge of interest, there are major gaps in our understanding of the fundamentals of neural net models. Understanding the mechanism behind their extraordinary generalization properties remains an open problem. A significant challenge arises in the non-convexity of training DNNs. In non-convex optimization, the choice of optimization method and its internal parameters such as initialization, mini-batching and step sizes have a considerable effect on the quality of the learned model. This is in sharp contrast to convex optimization problems, where these optimization parameters have no effect, and globally optimal solutions can be obtained in a very robust, efficient, transparent and reproducible manner.

In this talk, we introduce exact convex optimization formulations of multilayer neural network training problems. We show that two and three layer neural networks with ReLU or polynomial activations can be globally trained via convex programs with the number of variables polynomial in the number of training samples and number of hidden neurons. Our results provide an equivalent characterization of neural networks as convex models where a mixture of locally linear models are fitted to the data with sparsity inducing convex regularization. Moreover, we show that certain standard two and three layer convolutional neural networks can be globally optimized in fully polynomial time. We discuss extensions to batch normalization and generative adversarial networks. Finally, we present numerical simulations verifying our claims and illustrating that standard local search heuristics such as stochastic gradient descent can be inefficient compared to the proposed convex program.

Date and Time: 
Thursday, June 3, 2021 - 4:30pm

ISL Colloquium: Multi-Scale Methods for Machine Learning

Topic: 
Multi-Scale Methods for Machine Learning
Abstract / Description: 

Many modern applications require us to very quickly find relevant results from an enormous output space of potential candidates, for example, finding the best matching product from a large catalog or suggesting related search phrases on a search engine. The size of the output space for these problems can be in the millions to billions. Moreover, observational or training data is often limited for many of the so-called "long-tail" of items in the output space. Given the inherent paucity of training data for most of the items in the output space, developing machine learning models that perform well for spaces of this size is a contemporary challenge. In this talk, I will present a multi-scale machine learning framework called Prediction for Enormous and Correlated Output Spaces (PECOS). PECOS proceeds by first building a hierarchy over the output space using unsupervised learning, and then learning a machine learning model that makes predictions at each level of the hierarchy. Finally, the multi-scale predictions are combined to obtain an overall predictive model. This leads to an inference method that scales logarithmically with the size of the output space. A key to obtaining high performance is leveraging sparsity and developing highly efficient sparse matrix routines.

Date and Time: 
Thursday, May 20, 2021 - 4:30pm

ISL Colloquium: Interior Point Methods for Nearly Linear Time Algorithms

Topic: 
Interior Point Methods for Nearly Linear Time Algorithms
Abstract / Description: 

Linear programming is a foundational continuous optimization problem and special cases, e.g. maximum cardinality bipartite matching, are among the most fundamental problems in combinatorial optimization. In this talk, I will survey recent advances in solving these problems using interior point methods. I will discuss how new interior point methods can be leveraged to efficiently reduce these problems to data structures problems which can be solved efficiently with techniques from randomized numerical linear algebra and graph theory. This talk will highlight recent joint work which showed that linear programs with d variables and n constraints can be solved with high probability to high precision in time Otilde(nd + poly(d)) and that a variety of graph problems, e.g. maximum flow and optimal transport, on n-node -m-edge graphs can be solved to with high probability to high precision in time Otilde(m + n^1.5). These results constitute the first polynomial time methods for solving these problems to high precision which run in nearly-linear time on sufficiently large and dense instances. No previous experience with interior point methods required.
This talk will touch upon joint work with Jan van den Brand, Yin Tat Lee, Yang P. Liu, Danupon Nanongkai, Richard Peng, Thatchaphol Saranurak, Zhao Song, and Di Wang including 2009.01802, 2002.02304, and 2101.05719.

Date and Time: 
Thursday, May 6, 2021 - 4:30pm

ISL Colloquium: Private Stochastic Convex Optimization

Topic: 
Private Stochastic Convex Optimization
Abstract / Description: 

I will summarize some recent works on differentially private (DP) algorithms for stochastic convex optimization: the problem of minimizing the population loss given i.i.d. samples from a distribution over convex loss functions. In the standard l2/l2 setting, we will see two approaches to getting optimal rates for this problem. We show that for a wide range of parameters, privacy causes no additional overhead in accuracy or run time. In the process, we will develop techniques for private stochastic optimization that work for other geometries. For the LASSO setting when optimizing over the l1 ball, we will see private algorithms that achieve optimal rates. Based on joint works with various subsets of Hilal Asi, Raef Bassily, Vitaly Feldman, Tomer Koren and Abhradeep Thakurta.

Date and Time: 
Thursday, April 29, 2021 - 4:30pm

ISL Colloquium presents "Adaptive Experimental Design for Best Identification and Multiple Testing"

Topic: 
Recent results in planted assignment problems
Abstract / Description: 

Motivated by applications such as particle tracking, network de-anonymization, and computer vision, a recent thread of research is devoted to statistical models of assignment problems, in which the data are random weight graphs correlated with the latent permutation. In contrast to problems such as planted clique or stochastic block model, the major difference here is the lack of low-rank structures, which brings forth new challenges in both statistical analysis and algorithm design.

In the first half of the talk, we discuss the linear assignment problem, where the goal is to reconstruct a perfect matching planted in a randomly weighted bipartite graph, whose planted and unplanted edge weights are independently drawn from two different distributions. We determine the sharp threshold at which the optimal reconstruction error (fraction of misclassified edges) exhibits a phase transition from imperfect to perfect. Furthermore, for exponential weight distributions, this phase transition is shown to be of infinite-order, confirming the conjecture in [Semerjian et al. 2020]. The negative result is shown by proving that, below the threshold, the posterior distribution is concentrated away from the hidden matching by constructing exponentially many long augmenting cycles.

In the second half of the talk, we discuss the quadratic assignment problem (graph matching), where the goal is to recover the hidden vertex correspondence between two edge-correlated Erdos-Renyi graphs. We prove that there exists a sharp threshold, above which one can correctly match all but a vanishing fraction of the vertices and below which matching any positive fraction is impossible, a phenomenon known as the "all-or-nothing" phase transition. The proof builds upon a tight characterization of the mutual information via the truncated second-moment method and an appropriate "area theorem". Achieving these thresholds with efficient algorithms remains open.

This talk is based on joint work with Jian Ding, Jiaming Xu, Dana Yang and Sophie Yu. Preprints available at: https://arxiv.org/abs/2103.09383, https://arxiv.org/abs/2008.10097, https://arxiv.org/abs/2102.00082.

Date and Time: 
Thursday, April 22, 2021 - 4:30pm

ISL Colloquium presents "Demystifying the Efficiency of Reinforcement Learning: Two Recent Stories"

Topic: 
Demystifying the Efficiency of Reinforcement Learning: Two Recent Stories
Abstract / Description: 

Reinforcement learning (RL), which is frequently modeled as sequential learning and decision making in the face of uncertainty, is garnering growing interest in recent years due to its remarkable success in practice. In contemporary RL applications, it is increasingly more common to encounter environments with prohibitively large state and action space, thus imposing stringent requirements on the sample and computational efficiency of the RL algorithms in use. Despite the empirical success, however, the theoretical underpinnings for many popular RL algorithms remain highly inadequate even for the tabular setting.

In this talk, we present two vignettes regarding the effectiveness of RL algorithms. The first vignette demonstrates that a perturbed model-based RL approach is minimax optimal under a generative model, without suffering from a sample size barrier that was present in all past work. The second vignette covers policy optimization in reinforcement learning. On the one hand, we demonstrate that the popular softmax policy gradient method can take exponential time to converge; on the other hand, employing natural policy gradients and enforcing entropy regularization provably achieve fast global convergence. These results cover two distinctive RL paradigms, and might shed light on the efficacy of these algorithms in more complicated scenarios.

Date and Time: 
Thursday, April 15, 2021 - 4:30pm

Reinforcement Learning Forum Talk: Provable Model-based Nonlinear Bandit and Reinforcement Learning

Topic: 
Provable Model-based Nonlinear Bandit and Reinforcement Learning
Abstract / Description: 

Deep model-based reinforcement learning methods have achieved state-of-the-art sample efficiency but we lack a theoretical understanding of them. This talk will first show that convergence to a global maximum requires an exponential number of samples even for a one-layer neural net bandit problem, which is strictly easier than RL. Therefore, we propose to study convergence to local maxima. For both nonlinear bandit and RL, I will present a model-based algorithm, Virtual Ascent with Online Model Learner (ViOL), which provably converges to a local maximum with sample complexity that only depends on the sequential Rademacher complexity of the model class. Our results imply novel global or local regret bounds on several concrete settings such as linear bandit with finite or sparse model class, and two-layer neural net bandit.

Paper: https://arxiv.org/pdf/2102.04168.pdf


Details available at RL site, link below

 

Date and Time: 
Thursday, April 15, 2021 - 2:30pm

ISL Colloquium presents "Adaptive Experimental Design for Best Identification and Multiple Testing"

Topic: 
Scalable semidefinite programming
Abstract / Description: 

Semidefinite programming (SDP) is a powerful framework from convex optimization that has striking potential for data science applications. This talk describes a provably correct randomized algorithm for solving large, weakly constrained SDP problems by economizing on the storage and arithmetic costs. Numerical evidence shows that the method is effective for a range of applications, including relaxations of MaxCut, abstract phase retrieval, and quadratic assignment problems. Running on a laptop equivalent, the algorithm can handle SDP instances where the matrix variable has over 10^14 entries.

This talk will highlight the ideas behind the algorithm in a streamlined setting. The insights include a careful problem formulation, design of a bespoke optimization method, use of randomized eigenvalue computations, and use of randomized sketching methods.

Joint work with Alp Yurtsever, Olivier Fercoq, Madeleine Udell, and Volkan Cevher. Based on arXiv 1912.02949 (Scalable SDP, SIMODS 2021) and other papers (SketchyCGM in AISTATS 2017, Nyström sketch in NeurIPS 2017).

Date and Time: 
Thursday, April 8, 2021 - 4:30pm

Pages

IT-Forum

IT-Forum: "Between tractable and intractable problems in reinforcement learning"

Topic: 
Between tractable and intractable problems in reinforcement learning
Abstract / Description: 

Markov decision processes (MDPs) capture some of the most important aspects of decision making under uncertainty and as such they are at the heart of many efforts to decision making under uncertainty. However, MDPs are "flat" with no structure and as such, planning and learning in MDPs with multidimensional state spaces, common in applications, is provably intractable. Yet, reinforcement learning methods have been quite successful in providing strong solutions to some of these seemingly intractable problems. In this talk I will present my view of how to think about these successes by presenting a framework where the key idea is to give algorithms hints that can create backdoors to crack otherwise intractable problems. The talk will then dive into categorizing hints based on whether they can indeed succeed at doing this for the special case when the hints are given in the form of constraints on how value functions look like in the context of planning with generative models, also known as simulation optimization. As we shall see, seemingly minor differences between hints can cause some hints to work, while others fail.

Date and Time: 
Thursday, June 10, 2021 - 4:30pm

IT-Forum: "The Gittins Policy is Nearly Optimal in the M/G/k under Extremely General Conditions"

Topic: 
The Gittins Policy is Nearly Optimal in the M/G/k under Extremely General Conditions
Abstract / Description: 

The Gittins scheduling policy minimizes the mean response in the single-server M/G/1 queue in a wide variety of settings. Most famously, Gittins is optimal when preemption is allowed and service requirements are unknown but drawn from a known distribution. Gittins is also optimal much more generally, adapting to any amount of available information and any preemption restrictions. However, scheduling to minimize mean response time in a multiserver setting, specifically the central-queue M/G/k, is a much more difficult problem.

In this work we give the first general analysis of Gittins in the M/G/k. Specifically, we show that under extremely general conditions, Gittins's mean response time in the M/G/k is at most its mean response time in the M/G/1 plus a term that grows logarithmically in the heavy-traffic limit, namely as the system load approaches capacity. A consequence of this result is that Gittins is optimal in the heavy-traffic M/G/k if the service requirement distribution has finite (2 + ε)th moment for any ε > 0. This is the most general result on minimizing mean response time in the M/G/k to date.

To prove our results, we combine properties of the Gittins policy and Palm calculus in a novel way. Notably, our technique overcomes the limitations of tagged job methods that were used in prior scheduling analyses.

Joint work with Isaac Grosof and Mor Harchol-Balter. Paper link: https://ziv.codes/publications/#the-gittins-policy-is-nearly-optimal-in-the-mgk-under-extremely-general-conditions

Date and Time: 
Friday, May 21, 2021 - 1:00pm to Saturday, May 22, 2021 - 12:55pm

Algorithms & Friends Seminar presents "Modeling the Heterogeneity in COVID-19’s Reproductive Number and Its Impact on Predictive Scenarios"

Topic: 
Modeling the Heterogeneity in COVID-19’s Reproductive Number and Its Impact on Predictive Scenarios
Abstract / Description: 

The correct evaluation of the reproductive number R for COVID-19 — which characterizes the average number of secondary cases generated by each typical primary case— is central in the quantification of the potential scope of the pandemic and the selection of an appropriate course of action. In most models, R is modelled as a universal constant for the virus across outbreak clusters and individuals— effectively averaging out the inherent variability of the transmission process due to varying individual contact rates, population densities, demographics, or temporal factors amongst many. Yet, due to the exponential nature of epidemic growth, the error due to this simplification can be rapidly amplified and lead to inaccurate predictions and/or risk evaluation. From the statistical modeling perspective, the magnitude of the impact of this averaging remains an open question: how can this intrinsic variability be percolated into epidemic models, and how can its impact on uncertainty quantification and predictive scenarios be better quantified? In this talk, we discuss a Bayesian perspective on this question, creating a bridge between the agent-based and compartmental approaches commonly used in the literature. After deriving a Bayesian model that captures at scale the heterogeneity of a population and environmental conditions, we simulate the spread of the epidemic as well as the impact of different social distancing strategies, and highlight the strong impact of this added variability on the reported results. We base our discussion on both synthetic experiments — thereby quantifying of the reliability and the magnitude of the effects — and real COVID-19 data.


 

Hosted by the Algorithms and Friends Seminar

Date and Time: 
Monday, April 5, 2021 - 12:00pm to Tuesday, April 6, 2021 - 11:55am

ISL Colloquium presents "Finding Global Minima via Kernel Approximations"

Topic: 
Finding Global Minima via Kernel Approximations
Abstract / Description: 

We consider the global minimization of smooth functions based solely on function evaluations. Algorithms that achieve the optimal number of function evaluations for a given precision level typically rely on explicitly constructing an approximation of the function which is then minimized with algorithms that have exponential running-time complexity. In this paper, we consider an approach that jointly models the function to approximate and finds a global minimum. This is done by using infinite sums of square smooth functions and has strong links with polynomial sum-of-squares hierarchies. Leveraging recent representation properties of reproducing kernel Hilbert spaces, the infinite-dimensional optimization problem can be solved by subsampling in time polynomial in the number of function evaluations, and with theoretical guarantees on the obtained minimum. (Joint work with Alessandro Rudi and Ulysse Marteau-Ferey).

Date and Time: 
Thursday, April 1, 2021 - 10:00am to Friday, April 2, 2021 - 9:55am

IT-Forum presents "Approximating cross-validation: guarantees for model assessment and selection"

Topic: 
Approximating cross-validation: guarantees for model assessment and selection
Abstract / Description: 

Cross-validation (CV) is the de facto standard for selecting accurate predictive models and assessing model performance. However, CV suffers from a need to repeatedly refit a learning procedure on a large number of training datasets. To reduce the computational burden, a number of works have introduced approximate CV procedures that simultaneously reduce runtime and provide model assessments comparable to CV when the prediction problem is sufficiently smooth. An open question however is whether these procedures are suitable for model selection. In this talk, I'll describe (i) broad conditions under which the model selection performance of approximate CV nearly matches that of CV, (ii) examples of prediction problems where approximate CV selection fails to mimic CV selection, and (iii) an extension of these results and the approximate CV framework more broadly to non-smooth prediction problems like L1-regularized empirical risk minimization.

Date and Time: 
Thursday, March 18, 2021 - 4:30pm
Venue: 
Zoom: register to receive ID

IT Forum presents "Robust Learning from Batches – The Best Things in Life are (Almost) Free"

Topic: 
Robust Learning from Batches – The Best Things in Life are (Almost) Free
Abstract / Description: 

In many applications, including natural language processing, sensor networks, collaborative filtering, and federated learning, data are collected in batches, some potentially corrupt, biased, or even adversarial. Learning algorithms for this setting have therefore garnered considerable recent attention. We develop a general framework for robust learning from batches, and determine the least number of samples required for robust density estimation and classification over both discrete and continuous domains. Perhaps surprisingly, we show that robust learning can be achieved with essentially the same number of samples as required for genuine data. For the important problems of learning discrete and piecewise-polynomial densities, and of interval-based classification, we achieve these limits with polynomial-time algorithms.

Based on joint work with Ayush Jain.

Date and Time: 
Thursday, March 11, 2021 - 4:30pm

Algorithms & Friends Seminar presents "How fast do algorithms improve?"

Topic: 
How fast do algorithms improve?
Abstract / Description: 

Algorithms are one of the fundamental building blocks of computing. But current evidence about how fast algorithms improve is anecdotal, using small numbers of case studies to extrapolate. In this work, we gather data from 57 textbooks and more than 1,137 research papers to present the first systematic view of algorithm progress ever assembled. There is enormous variation. Around half of all algorithm families experience little or no improvement. At the other extreme, 13% experience transformative improvements, radically changing how and where they can be used. Overall, we find that, for moderate-sized problems, 30% to 45% of algorithmic families had improvements comparable or greater than those that users experienced from Moore's Law and other hardware advances.

Joint work with Yash M. Sherry.

Date and Time: 
Monday, February 22, 2021 - 12:00pm to Tuesday, February 23, 2021 - 11:55am

IT Forum presents "Distribution-Free, Risk-Controlling Prediction Sets"

Topic: 
Distribution-Free, Risk-Controlling Prediction Sets
Abstract / Description: 

To communicate instance-wise uncertainty for prediction tasks, we show how to generate set-valued predictions for black-box predictors that control the expected loss on future test points at a user-specified level. Our approach provides explicit finite-sample guarantees for any dataset by using a holdout set to calibrate the size of the prediction sets. This framework enables simple, distribution-free, rigorous error control for many tasks, and we demonstrate it in five large-scale machine learning problems: (1) classification problems where some mistakes are more costly than others; (2) multi-label classification, where each observation has multiple associated labels; (3) classification problems where the labels have a hierarchical structure; (4) image segmentation, where we wish to predict a set of pixels containing an object of interest; and (5) protein structure prediction.

Date and Time: 
Friday, February 19, 2021 - 1:00pm to Saturday, February 20, 2021 - 12:55pm

IT Forum presents "Scaling Wasserstein distances to high dimensions via smoothing"

Topic: 
Scaling Wasserstein distances to high dimensions via smoothing
Abstract / Description: 

Wasserstein distances has recently seen a surge of applications in statistics and machine learning. This stems from many advantageous properties they possess, such as metric structure (they metrize weak convergence), robustness to support mismatch, compatibility to gradient-based optimization, and rich geometric properties. In practice, we rarely have access to the actual distribution and only get data from it, which necessitates estimating the distance from samples. A central issue is that such estimators suffer from the curse of dimensionality: their empirical convergence rate scales as n^{-1/d} for d-dimensional distributions. This rate deteriorates exponentially fast with dimension, making it impossible to obtain meaningful accuracy guarantees, considering the dimensionality of real-world data.

This talk will present a novel framework of smooth Wasserstein distances, that inherits the properties of their classic counterparts while alleviating the empirical curse of dimensionality. Specifically, we will show that the empirical approximation error of the smooth distance decays as n^{-1/2}, in all dimensions. For the special case of the smooth 1-Wasserstein distance, we will also derive a high-dimensional limit distribution, further highlighting the favorable statistical behavior of the smooth framework. Applications to implicit generative modeling will be considered, leveraging the statistical efficiency results to establish n^{-1/2} generalization bounds in any dimension.

Date and Time: 
Friday, February 12, 2021 - 1:00pm to Saturday, February 13, 2021 - 12:55pm

Algorithms & Friends Seminar presents "A New Approach for Fighting Infectious Disease, Combining Game Theory and Network Theory"

Topic: 
A New Approach for Fighting Infectious Disease, Combining Game Theory and Network Theory
Abstract / Description: 

There is a categorically new way to fight infectious disease, which could have a significant impact on the COVID pandemic now. Its origins come from game theory and computer science. It works against every COVID variant, and would play a key role in mopping up vaccine-resistant infections to block COVID's evolution. It's an app which is fundamentally different from every other app (and which resolves deep flaws in "contact tracing apps"). Functionally, it gives you an anonymous radar that tells you how "far" away COVID has just struck. "Far" is measured by counting physical relationships (https://novid.org, https://youtu.be/EIU-6FvwikQ). The simple idea flips the incentives. Previous approaches were about controlling you, preemptively removing you from society if you were suspected of being infected. This new tool lets you see incoming disease to defend yourself just in time. This uniquely aligns incentives so that even if everyone in a democratic society does what is best for themselves, they end up doing what is best for the whole.

Date and Time: 
Monday, February 8, 2021 - 12:00pm to Tuesday, February 9, 2021 - 11:55am
Venue: 
Registration required

Pages

Optics and Electronics Seminar

SPRC & SOS Seminar: Trace Gas Flux Measurements in Artic Ecosystems – Current Applications and Future Perspectives

Topic: 
Trace Gas Flux Measurements in Artic Ecosystems – Current Applications and Future Perspectives
Abstract / Description: 

The exchanges of carbon dioxide (CO2) and methane (CH4) between arctic ecosystems and the atmosphere are sizeable and may if they change may alter the further development of climate warming. These fluxes take place between the atmospheric background of these trace gases and a very large reservoir of organic carbon stored in arctic soils and sediments. The stored carbon amounts to more than twice the current global atmospheric burden of CO2. Therefore, accurate measurements of the fluxes and how they vary temporally and spatially is key to understanding possible ecosystem impacts on the atmospheric mixing ratios and hence climate.
This presentation will detail the current state of the global CO2 and CH4 budget and review methods applied in quantifying the arctic trace gas flux components. The presentation will show new developments in the attempts to use UAV's to help improve spatial coverage of measured fluxes. The challenges in applying the laser techniques needed for the precise and fast trace gas measurements under true arctic conditions will be discussed.

Date and Time: 
Tuesday, May 4, 2021 - 11:00am to Wednesday, May 5, 2021 - 10:55am

Optics & Electronics seminar: Inverse Designed Integrated Photonics

Topic: 
[AP483] Inverse Designed Integrated Photonics
Abstract / Description: 

Photonics with superior properties can be implemented in a variety of old (silicon, silicon nitride) and new (silicon carbide, diamond) photonic materials by combining state-of-the- art photonics optimization techniques (photonics inverse design) with new fabrication approaches. In addition to making photonics more robust (e.g., to errors in fabrication and variation in temperature), more compact, and more efficient, this approach can also enable new functionalities. While in our early work we focused on inverse design and demonstration of individual photonic devices, our more recent work focused on scaling it to photonic integrated circuits fabricated in a commercial semiconductor foundry. We illustrate this with several examples, including optical interconnects based on a combination of mode and wavelength division multiplexing, and photonic chips for optical beam steering.


 

Spring 2021 speakers organized by Prof. David A.B. Miller

Date and Time: 
Monday, May 24, 2021 - 4:15pm

Optics & Electronics seminar: Energy Efficient, Multi-Terabit Photonic Connectivity for Disaggregated Computing

Topic: 
[AP483] Energy Efficient, Multi-Terabit Photonic Connectivity for Disaggregated Computing
Abstract / Description: 

High performance data centers are increasingly bottlenecked by the energy and communications costs of interconnection networks. Integrated silicon photonics with dense wavelength-division multiplexing offer the opportunity of optical connectivity that directly delivers high off-chip communication bandwidth densities with low power consumption. We introduce the concept of embedded photonics for deeply disaggregated architectures. Beyond alleviating the bandwidth/energy bottlenecks, the new architectural approach enables flexible connectivity tailored for specific applications.


 

Spring 2021 speakers organized by Prof. David A.B. Miller

Date and Time: 
Monday, May 3, 2021 - 4:15pm

Optics & Electronics seminar: End to End Optimization of Optics and Algorithms

Topic: 
[AP483] End to End Optimization of Optics and Algorithms
Abstract / Description: 

Neural networks and other advanced image processing algorithms excel in a wide variety of computer vision and imaging applications, but their high performance also comes at a high computational cost, and their success is sometimes limited. In this talk, we explore hybrid optical-digital strategies to computational imaging that outsource parts of the algorithm into the optical domain. Using such a co-design of optics and image processing, we can design application-domain-specific cameras or compute parts of a convolutional neural network in optics. Optical computing happens at the speed of light and without any memory or power requirements, thereby opening new directions for intelligent imaging systems.


 

Spring 2021 speakers organized by Prof. David A.B. Miller

Date and Time: 
Monday, April 26, 2021 - 4:15pm

Optics & Electronics seminar: Playing Hide-and-Seek at the Nanoscale – Tailored Field Landscapes for Precise Particle Tracking and More

Topic: 
[AP483] Playing Hide-and-Seek at the Nanoscale – Tailored Field Landscapes for Precise Particle Tracking and More
Abstract / Description: 

Intensity, phase, and polarization of electromagnetic fields can be tailored to form bespoke distributions. This spatial structure of light is an intriguing playground. It finds a wide range of applications in optical communications, sensing, imaging, and beyond. However, the real beauty of structured fields comes to light when considering strongly confined electromagnetic fields, which naturally feature complex three-dimensional distributions. These engineered light fields can be utilized, for example, to selectively excite individual nanosystems, extending the range of possible applications even further by enabling advanced single-particle spectroscopy, nanoscale traffic control, nano-metrology, and more.

In this talk, we introduce a novel scheme for ultra-precise particle localization and tracking, which is based on the interaction of structured light and individual nanoparticles. We show that by utilizing the strongly position-dependent scattering pattern as a ruler, we can retrieve the exact particle position from the far-field scattering signal. Besides the fundamental ingredients and recipe of this technique, we also highlight selected next steps and future plans. For instance, we also discuss the fundamental idea behind next generation camera technology to boost the capabilities of our nanometrology schemes. These novel pixels and cameras based on light-processing circuitry are jointly developed, together with other European and international experts, as part of the 'SuperPixels' project (https://www.superpixels.org/). They will allow for a simultaneous measurement of polarization and phase distributions in addition to light's intensity.

[This project has received partial funding from the European Commission Horizon 2020 research and innovation program under the Future and Emerging Technologies Open Grant Agreement Super-Pixels No. 829116]


 

Spring 2021 speakers organized by Prof. David A.B. Miller

Date and Time: 
Monday, April 19, 2021 - 12:00pm to Tuesday, April 20, 2021 - 11:55am

Optics & Electronics seminar: Semiconductor Nanocrystal Optoelectronics: A Journey from Colloidal Quantum Dots to Wells

Topic: 
[AP483] Semiconductor Nanocrystal Optoelectronics: A Journey from Colloidal Quantum Dots to Wells
Abstract / Description: 

Semiconductor nanocrystals have attracted great interest for color conversion and enrichment in quality lighting and displays. Optical properties of these solution‑processed nanostructures are conveniently controlled by tailoring their size, shape, and composition in an effort to realize high‑performance light generation and lasing. These colloids span different types and heterostructures of semiconductors in the forms of quantum dots and rods to the latest sub‑family of nanocrystals, the colloidal quantum wells (CQWs). In this talk, we will introduce the emerging field of semiconductor nanocrystal optoelectronics, with most recent examples of their photonic structures and optoelectronic devices employing such atomically‑flat, tightly‑confined, quasi‑two‑dimensional CQWs, also popularly nick-named 'nanoplatelets'. Among their various extraordinary features, we will show that these CQWs enable record high optical gain coefficients [1] and can achieve gain thresholds at the level of sub‑single exciton population per CQW on average [2], empowered by carefully engineering their heterostructure [3]. Next, we will present a new, powerful, large-area self‑assembly tool for orientation controlling of these nanoplatelets [4], which provides us with the ability to tune and master their excitonic properties in their ensemble, as well as the level of achievable energy transfer among them and with other nearby species. Using three‑dimensional constructs of face‑down self‑assembled slabs of CQWs with monolayer precision, we will demonstrate ultrathin optical gain media and lasers of these oriented CQW assemblies [5]. Finally, we will show record high‑efficiency colloidal LEDs using CQWs employed as the electrically‑driven active emitter layer [6] and record low-threshold solution lasers using the same CQWs employed as the optically‑pumped fluidic gain medium [7]. Given their current accelerating progress, these solution‑processed quantum well materials hold great promise to challenge their epitaxial thin‑film counterparts in semiconductor optoelectronics in the near future.

[1] B. Guzelturk et al., HVD, Nano Letters 19, 277 (2019)
[2] N. Taghipour et al., HVD, Nature Comm 11, 3305 (2020)
[3] Y. Altıntas et al., HVD, ACS Nano 13, 10662 (2019)
[4] O. Erdem et al., HVD, Nano Letters 19, 4297 (2019)
[5] O. Erdem et al., HVD, Nano Letters 20, 6459 (2020)
[6] B. Liu et al., HVD, Advanced Materials 32, 1905824 (2020)
[7] J. Maskoun et al., HVD, Advanced Materials (2021), in print, DOI:(10.1002/adma.202007131)


Spring 2021 speakers organized by Prof. David A.B. Miller

Date and Time: 
Monday, April 12, 2021 - 12:00pm to Tuesday, April 13, 2021 - 11:55am

Optics & Electronics seminar: Creation of Time Reversed Optical Waves

Topic: 
[AP483] Creation of Time Reversed Optical Waves
Abstract / Description: 

In this talk, a new type of beam shaper will be discussed, capable of generating arbitrary vector spatiotemporal beams, where the user can define the amplitude, phase, and polarization independently for each point in space and time. This beam shaper was recently used to demonstrate time reversed optical waves. Such waves propagate through complex media, as if watching a traditional scattering process in reverse - starting as a complicated ‘pre-scattered’ wave, which then becomes a desired target field at the distal end of the complex media.


Spring 2021 speakers organized by Prof. David A.B. Miller

 

Date and Time: 
Monday, April 5, 2021 - 4:15pm

EST Seminar presents "Oxide Semiconductor Electronics: from BEOL thin film circuits to high-voltage UWBG"

Topic: 
Oxide Semiconductor Electronics: from BEOL thin film circuits to high-voltage UWBG
Abstract / Description: 

Oxide semiconductors' unique properties – a wide bandgap, reasonably high electron mobility, and ease of bulk and thin film preparation – make them prime material candidates for a variety of electronic devices. In this talk, I will describe my group's recent work on amorphous and crystalline oxide semiconductors for back end of line (BEOL) thin film circuitry and for high-voltage ultra-wide bandgap (UWBG) power devices, respectively. First, we exploit the thermodynamics of interfacialin situredox reactions with amorphous zinc tin oxide (a-ZTO) semiconductor to make MISFETs, MESFETs, Schottky diodes, and resistive memory devices that can be monolithically integrated with silicon CMOS. We demonstrate rectifiers that can harvest RFID wireless power and inverters that are compatible with low voltage silicon ICs. Furthermore, we develop novel, scalable atomic layer deposition processes to realize high-qualitya-ZTO semiconductor films, high-kAl2O3gate insulators and passivation layers, and Al:ZnO source/drain electrodes. Combining an innovative electro-hydro-dynamic jetting process with additive and subtractive selective area ALD, we realize thin film transistors with channel lengths below the ink-jet printing limit. I'll also describe our recent work on p-type oxide TFTs using Cu2O, and explain the key challenges in device architecture and materials physics that limit p-TFT performance. Finally, I'll explain how we've taken the learning from thin film oxides and used it to realize ultra-stable ohmic contacts and MOS capacitors to crystalline beta-phase gallium oxide, an ultra-wide bandgap semiconductor of interest for multi-kV power devices.

Date and Time: 
Tuesday, March 30, 2021 - 1:00pm to Wednesday, March 31, 2021 - 12:55pm

Mapping urban emissions with neighborhood resolution

Topic: 
Mapping urban emissions with neighborhood resolution
Abstract / Description: 

Most people on earth live in cities and they are responsible for the majority of greenhouse gas emissions. Cities are also where exposure to poor air quality is most frequent and most variable. Understanding and managing the path to net-zero greenhouse gas emissions, improved public health and lower public health inequities requires a view into the emissions and atmospheric chemistry of cities with the fine-grained detail that allows evaluation of specific processes and variations from one neighborhood to another. In this talk, I will describe the development of the Berkeley Environment, Air Quality and CO2 Network (BEACO2N http://beacon.berkeley.edu/about/), a dense network for mapping urban CO2, NOx, CO, O3 and aerosol. Each node of the network contains multiple optical and ChemFET sensors which accurately quantify the trace gases and are calibrated in the field using a variety of cross-referencing methods and comparisons with traceable standards. Integration of the BEACO2N maps with simple Gaussian plume models and sophisticated inversions employing high resolution weather models provide unique observational constraints on spatial and temporal patterns of CO2 and other emissions. Examples from the COVID shelter-in-place and testing models of fuel efficiency vs. vehicle speed in the SF Bay Area will be described along with prospects for further extension to other cities and other chemicals.



Organized by Stanford OSA/SPIE Student Chapter with SPRC and Ginzton Lab

Date and Time: 
Tuesday, March 9, 2021 - 1:30pm

Pages

SCIEN Talk

SCIEN Colloquium and EE 292E presents "Learning-based 3D Computer-Generated Holography"

Topic: 
Learning-based 3D Computer-Generated Holography
Abstract / Description: 

Computer generated holography (CGH) is fundamental to applications such as biosensing, volumetric display, optical/acoustic tweezer, security and many others that require spatial control of intricate optical or acoustic fields. For near-eye displays, CGH provides the opportunity to support true 3D projection in a sunglass-like display. Yet, the conventional approach to compute a true 3D hologram via physical simulation of diffraction and inference is slow and unaware of occlusion. Moreover, experimental results are often inferior to simulations due to non-idealized optical systems, non-linear and non-uniform SLM responses, and image degradation caused by complex to phase-only conversion. These computational and hardware-imposed challenges together limit the interactiveness and realism of the ultimate immersive experience. In this talk, I will describe techniques to mitigate these challenges, including physical simulation algorithms that handle occlusion for RGB-D and more advanced 3D input, methods to create large-scale 3D hologram datasets, training of CNNs to speed up complex and phase-only hologram synthesis, and approaches to compensate hardware limitations. Together, the resulted system can synthesis and display photorealistic 3D holograms in real-time using a single consumer-grade GPU and run interactively on an iPhone leveraging the Neural Engine. I will further discuss possible extensions that could be built top of the proposed system to support foveated rendering, static pupil expansion, view-dependent effect and other features.

Date and Time: 
Wednesday, June 2, 2021 - 4:30pm

SCIEN and EE292E presents "Real-Time Ray Tracing and the Reinvention of the Graphics Pipeline"

Topic: 
Real-Time Ray Tracing and the Reinvention of the Graphics Pipeline
Abstract / Description: 

For many years, real-time ray tracing was the technology of the future; in 2008, David Kirk famously quipped that it always would be. There were plenty of reasons to doubt that the approach would be suitable for real-time rendering, many of them firmly believed by the speaker. Yet dedicated hardware for ray tracing has now arrived in recent GPUs. Its greatest successes so far have come from not the direct application of existing offline ray-tracing algorithms to real-time, but instead from the reinvention of fundamental rendering algorithms accounting for the constraints of real-time rendering. In this talk, I will survey the history of real-time ray tracing and some of the near misses along the way. I'll then discuss how real-time rendering is changing with the capabilities offered by the high-performance and arbitrary visibility queries offered by ray tracing.

Date and Time: 
Wednesday, May 26, 2021 - 4:30pm

SCIEN and EE292E present "Mantis shrimp–inspired organic photodetector"

Topic: 
Mantis shrimp–inspired organic photodetector for simultaneous hyperspectral and polarimetric imaging-enabling advanced single-pixel architectures
Abstract / Description: 

Many spectral and polarimetric cameras implement complex spatial, temporal, and spectral re-mapping strategies to measure a signal within a given use-case's specifications and error tolerances. This re-mapping results in a complex tradespace that is challenging to navigate; a tradespace driven, in part, by the limited degrees of freedom available in inorganic detector technology. This presentation overviews a new kind of organic detector and pixel architecture that enables single-pixel tandem detection of both spectrum and polarization. By using organic detectors' semitransparency and intrinsic anisotropy, the detector minimizes spatial and temporal resolution tradeoffs while showcasing thin-film polarization control strategies.

Date and Time: 
Wednesday, May 19, 2021 - 4:30pm

SCIEN and EE292E present "Photographic Image Priors in the Era of Machine Learning"

Topic: 
Photographic Image Priors in the Era of Machine Learning
Abstract / Description: 

Prior probability models are a central component of the statistical formulation of inverse problems, but density estimation is a notoriously difficult problem for high dimensional signals such as photographic images. Machine learning methods have produced impressive solutions for many inverse problems, greatly surpassing those achievable with simple prior models, but these are often not well understood and don't generalize well beyond their training context. About a decade ago, a new approach known as "plug-and-play" was proposed, in which a denoiser is used as an algorithmic component for imposing prior information. I'll describe our progress in understanding and using this implicit prior. We derive a surprisingly simple algorithm for drawing high-probability samples from the implicit prior embedded within a CNN trained to perform blind (i.e., unknown noise level) least-squares Gaussian denoising. A generalization of this algorithm to constrained sampling provides a method for solving *any* linear inverse problem, with no additional training, and no further distributional assumptions. We demonstrate this general form of transfer learning in multiple applications, using the same algorithm to produce state-of-the-art solutions for deblurring, super-resolution, and compressive sensing. I'll also discuss extensions to visualizing information capture in foveated visual systems. This is joint work with Zahra Kadkhodaie, Sreyas Mohan, and Carlos Fernandez-Granda

Date and Time: 
Wednesday, May 12, 2021 - 4:30pm

SCIEN colloquium and EE292E seminar: Pushing the Boundaries of Novel View Synthesis

Topic: 
Pushing the Boundaries of Novel View Synthesis
Abstract / Description: 

2020 was a turbulent year, but for 3D learning it was a fruitful one with lots of exciting new tools and ideas. In particular, there have been many exciting developments in the area of coordinate based neural networks and novel view synthesis. In this talk I will discuss our recent work on single image view synthesis with pixelNeRF, which aims to predict a Neural Radiance Field (NeRF) from a single image. I will discuss how NeRF representation allows models like pixel-aligned implicit functions (PiFu) to be trained without explicit 3D supervision and the importance of other key design factors such as predicting in view coordinate-frame and handling multi-view inputs. I will also touch upon our recent work that allows real-time rendering of NeRFs. Then, I will discuss Infinite Nature, a project in collaboration with teams at Google NYC, where we explore how to push the boundaries of novel view synthesis and generate views way beyond the edges of the initial input image, resulting in a controllable video generation of a natural scene.

Date and Time: 
Wednesday, May 5, 2021 - 4:30pm

SCIEN and EE292E present "Neural Representations: Coordinate Based Networks for Fitting Signals, Derivatives, and Integrals"

Topic: 
Neural Representations: Coordinate Based Networks for Fitting Signals, Derivatives, and Integrals
Abstract / Description: 

Implicitly defined, continuous, differentiable signal representations parameterized by neural networks have emerged as a powerful paradigm, offering many possible benefits over conventional representations. However, current network architectures for such implicit neural representations are incapable of modeling signals with fine detail, and fail to represent a signal's spatial and temporal derivatives, despite the fact that these are essential to many physical signals defined implicitly as the solution to partial differential equations. In this talk, we describe how sinusoidal representation networks or SIREN, are ideally suited for representing complex natural signals and their derivatives. Using SIREN, we demonstrate the representation of images, wavefields, video, sound, and their derivatives. Further, we show how SIRENs can be leveraged to solve challenging boundary value problems, such as particular Eikonal equations (yielding signed distance functions), the Poisson equation, and the Helmholtz and wave equations. While SIREN can be used to fit signals and their derivatives, we also introduce a new framework for solving integral equations using implicit neural representation networks. Our automatic integration framework, AutoInt, enables the calculation of any definite integral with two evaluations of a neural network. We apply our approach for efficient integration to the problem of neural volume rendering. Finally we present a novel architecture and training procedure able to fit data such as gigapixel images or fine-detailed 3D geometry, demonstrating those neural representations are now ready to be used in large scale scenarios.

Date and Time: 
Wednesday, April 28, 2021 - 4:30pm

SCIEN and EE292E present "The Chromatic Pyramid of Visibility"

Topic: 
The Chromatic Pyramid of Visibility
Abstract / Description: 

A fundamental limit to human vision is our ability to sense variations in light intensity over space and time. These limits have been formalized in the spatio-temporal contrast sensitivity function, which is now a foundation of vision science. This function has also proven to be the foundation of much applied vision science, providing guidance on spatial and temporal resolution for modern imaging technology. The Pyramid of Visibility is a simplified model of the human spatio-temporal luminance contrast sensitivity function (Watson, Andrew B.; Ahumada, Albert J. 2016). It posits that log sensitivity is a linear function of spatial frequency, temporal frequency, and log mean luminance. It is valid only away from the spatiotemporal frequency origin. It has recently been extended to peripheral vision to define the Field of Contrast Sensitivity (Watson 2018). Though very useful in a range of applications, the pyramid would benefit from an extension to the chromatic domain. In this talk I will describe our efforts to develop this extension. Among the issues we address are the choice of color space, the definition of color contrast, and how to combine sensitivities among luminance and chromatic pyramids.

Watson, A. B. (2018). "The Field of View, the Field of Resolution, and the Field of Contrast Sensitivity." Journal of Perceptual Imaging 1(1): 10505-10501-10505-10511.
Watson, A. B. and A. J. Ahumada (2016). "The pyramid of visibility." Electronic Imaging 2016(16): 1-6.

Date and Time: 
Wednesday, April 21, 2021 - 4:30pm

SCIEN and EE292E present "Neural Implicit Representations for 3D Vision"

Topic: 
Neural Implicit Representations for 3D Vision
Abstract / Description: 

In this talk, I will show several recent results of my group on learning neural implicit 3D representations, departing from the traditional paradigm of representing 3D shapes explicitly using voxels, point clouds or meshes. Implicit representations have a small memory footprint and allow for modeling arbitrary 3D topologies at (theoretically) arbitrary resolution in continuous function space. I will show the ability and limitations of these approaches in the context of reconstructing 3D geometry, texture and motion. I will further demonstrate a technique for learning implicit 3D models using only 2D supervision through implicit differentiation of the level set constraint. Finally, I will demonstrate how implicit models can tackle large-scale reconstructions and introduce GRAF and GIRAFFE which are generative 3D models for neural radiance fields that are able to generate 3D consistent photo-realistic renderings from unstructured and unposed image collections.

Date and Time: 
Wednesday, April 14, 2021 - 11:00am to Thursday, April 15, 2021 - 10:55am

SCIEN and EE292E present "Slow Glass"

Topic: 
Slow Glass
Abstract / Description: 

Wouldn't it be fascinating to be in the same room as Abraham Lincoln, visit Thomas Edison in his laboratory, or step onto the streets of New York a hundred years ago? We explore this thought experiment, by tracing ideas from science fiction through antique stereographs to the latest work in generative adversarial networks (GANs) to step back in time to experience these historical people and places not in black and white, but much closer to how they really appeared. In the process, I'll present our latest work on Keystone Depth, and Time Travel Rephotography.

Date and Time: 
Wednesday, April 7, 2021 - 4:30pm

SCIEN and EE292E present "Photographic Forensic Identification"

Topic: 
Photographic Forensic Identification
Abstract / Description: 

Forensic DNA analysis has been critical in prosecuting crimes and overturning wrongful convictions. At the same time, other physical and digital forensic identification techniques, used to link a suspect to a crime scene, are plagued with problems of accuracy, reliability, and reproducibility. Flawed forensic science can have devastating consequences: the National Registry of Exonerations identified that flawed forensic techniques contribute to almost a quarter of wrongful convictions in the United States. I will describe our recent efforts to examine the reliability of two such photographic forensic identification techniques: (1) identification based on purported distinct patterns in clothing; and (2) identification based on measurements of height and weight.

Date and Time: 
Wednesday, March 31, 2021 - 4:30pm

Pages

SmartGrid

TomKat Center Spring 2021 Innovation Showcase

Topic: 
TomKat Center Spring 2021 Innovation Showcase
Abstract / Description: 

Join us for a one-hour webinar featuring three of our Innovation Transfer teams. They will talk about the sustainable energy problems they are tackling, the technologies underpinning their solutions, and new developments for their products.
April 21, 2021 @ 4pm Pacific

Presentations from:

  • Jayce Hafner and Sami Tellatin, @FarmRaise - The one-stop-shop for farm grants and loans, informed by and built for farmers.
  • Rito Sur, @Indrio Technologies - Laser based sensors for rapid monitoring of chemical concentrations and harmful emissions.
  • Armelle Coutant, @KitSwitch - Adaptive reuse of underutilized commercial buildings into affordable housing.

 


Details and more information on TomKat Center site

 

Date and Time: 
Wednesday, April 21, 2021 - 4:00pm to Thursday, April 22, 2021 - 3:55pm
Venue: 
REGISTRATION REQUIRED

Bits & Watts' Smart Grid Seminar: Electric Grid Security

Topic: 
Electric Grid Security
Abstract / Description: 

Bits & Watts’ Smart Grid Seminar provides experts from academia, startups, research institutes, and large corporations to familiarize seminar participants with current challenges and advances in grid data analytics, economics, market design, battery storage, electrified transportation, power electronics, renewable energy. The seminar is open to Stanford students, as a 1 unit seminar course (CEE 272T/EE292T), and to the broader community.

Date and Time: 
Thursday, May 13, 2021 - 2:30pm

Bits & Watts' Smart Grid Seminar: ERCOT Frequency Regulation Ancillary Services

Topic: 
ERCOT Frequency Regulation Ancillary Services
Abstract / Description: 

Bits & Watts’ Smart Grid Seminar provides experts from academia, startups, research institutes, and large corporations to familiarize seminar participants with current challenges and advances in grid data analytics, economics, market design, battery storage, electrified transportation, power electronics, renewable energy. The seminar is open to Stanford students, as a 1 unit seminar course (CEE 272T/EE292T), and to the broader community.

Date and Time: 
Thursday, April 22, 2021 - 2:30pm

Bits & Watts' Smart Grid Seminar: Large-scale Grid Testbed

Topic: 
Large-scale Grid Testbed
Abstract / Description: 

Bits & Watts' Smart Grid Seminar provides experts from academia, startups, research institutes, and large corporations to familiarize seminar participants with current challenges and advances in grid data analytics, economics, market design, battery storage, electrified transportation, power electronics, renewable energy. The seminar is open to Stanford students, as a 1 unit seminar course (CEE 272T/EE292T), and to the broader community.

Date and Time: 
Thursday, April 15, 2021 - 2:30pm

Smart Grid Seminar: Dynamic Operation and Lifetime Valuation for Batteries in Markets

Topic: 
Dynamic Operation and Lifetime Valuation for Batteries in Markets
Abstract / Description: 

This talk introduces a generalized framework for valuating batteries based on their degradation mechanism and market signals. The proposed framework follows a dynamic programming approach, and novel solution algorithms are designed to accurately model years of battery degradation accumulation, while simulating high granularity optimal control for batteries over time resolutions of minutes or even seconds, including the capability to model stochastic operations. Results from this research will provide pricing and operation references for new and used batteries in grid-interactive applications, and demonstrate how batteries of different degradation mechanisms should be priced differently.

Date and Time: 
Thursday, November 19, 2020 - 1:30pm
Venue: 
Zoom

Smart Grid Seminar: IoT, Smart Microgrid

Topic: 
IoT, Smart Microgrid
Abstract / Description: 

Transforming the grid to a Smart Grid requires real time sensing, of energy producers, consumers, stored capacity in batteries, weather conditions, and more, essential to develop usage models, and anticipate need. Along with the ability to take 'action' such as bringing online cleaner sources of power and curbing demand we can help smooth load peaks. In this talk we introduce the open source project EdgeX Foundry and its viability to deliver a secure, extensible, endpoint to collect real time data, develop models, and control devices. Data processing at the edge while reducing network bandwidth needs, reduces response latency, and helps preserve data privacy. We illustrate this in the context of VMware's on-campus electric vehicle charging facility. We conclude by sharing our longer term goals of developing and controlling hierarchical composite device models and introduce open source project Kinney.

Seminars are open to all Stanford students, faculty, staff and community. Register via the RSVP link

Date and Time: 
Thursday, October 15, 2020 - 1:30pm
Venue: 
Zoom

Smart Grid Seminar: Increasing Electricity Market Efficiency with Flexible Mechanism

Topic: 
Increasing Electricity Market Efficiency with Flexible Mechanism
Abstract / Description: 

Current wholesale electricity market designs of day-ahead and real-time markets have inherent inefficiencies that are amplified with greater shares from intermittent renewable sources. Building in flexibility through time and geographically substitutable bids and stochastic unit commitments can yield significant gains. This talk will discuss how to model and assess these advantages and their implications for renewable generation.

Date and Time: 
Thursday, October 1, 2020 - 1:30pm
Venue: 
Zoom

SmartGrid Seminar presents "Cyber Security Research: A Data Scientist’s Perspective"

Topic: 
Cyber Security Research: A Data Scientist’s Perspective
Abstract / Description: 

Situational awareness of computer networks presents many challenges including but not limited to the volume of the data and the dynamic and evolving nature of the problem space. For example, at the perimeter of a corporate enterprise computer network, it is common to see terabytes of network traffic each day, containing millions of unique IP addresses and connection records that number in the hundreds of millions. Celeste will provide an overview and discuss recent trends facing computer security researchers and practitioners. She will describe recent work at Lawrence Livermore National Laboratory to enable analysis of computer networks and encourage an interactive discussion to foster new ideas to address these challenges.

Date and Time: 
Thursday, May 7, 2020 - 1:30pm
Venue: 
Register at tinyurl.com/Spring-SGS

Smart Grid Seminar presents "California Microgrids: Lessons Learned"

Topic: 
California Microgrids: Lessons Learned
Abstract / Description: 

There has been substantial discussion of the use of microgrids as a result of the recent wildfires and Public Safety Power Shutoffs. The ability of microgrids to be considered for these applications now, is in part a result of over a decade of research funded through the California Energy Commission's R&D programs. To date the CEC has issued over $100M in grants that have funded over 30 microgrids. These microgrids support a variety of applications, with a diversity of ownership models. This presentation will provide an overview of the microgrid research being funded by the CEC and future research areas of interest.

Date and Time: 
Thursday, April 23, 2020 - 1:30pm
Venue: 
Zoom: register to receive ID

Smart Grid Seminar presents "Grid Resilience"

Topic: 
Grid Resilience
Abstract / Description: 

The term "resilience" refers to the ability to prepare for and adapt to changing conditions and withstand and recover rapidly from disruptions. The value of resilience to the electricity system is in enabling the function of other interdependent infrastructure systems and maintaining critical energy-based services. Many resilience approaches are being considered. This webinar will outline some of these perspectives and provide an overview of some of the related ongoing activities within the Department of Energy.

Date and Time: 
Thursday, May 21, 2020 - 1:30pm
Venue: 
Zoom: register to receive ID

Pages

Stanford's NetSeminar

#StanfordToo: A Conversation about Sexual Harassment in Our Academic Spaces

Topic: 
#StanfordToo: A Conversation about Sexual Harassment in Our Academic Spaces
Abstract / Description: 

Individuals of all genders invited to be a part of:
#StanfordToo: A Conversation about Sexual Harassment in Our Academic Spaces, where we will feature real stories of harassment at Stanford academic STEM in a conversation with Provost Drell, Dean Minor (SoM), and Dean Graham (SE3). We will have plenty of time for audience discussion on how we can take concrete action to dismantle this culture and actively work towards a more inclusive Stanford for everyone. While our emphasis is on STEM fields, we welcome and encourage participation from students, postdocs, staff, and faculty of all academic disciplines and backgrounds.

Date and Time: 
Friday, April 19, 2019 - 3:30pm
Venue: 
STLC 111

John G. Linvill Distinguished Seminar on Electronic Systems Technology

Topic: 
Internet of Things and Internet of Energy for Connecting at Any Time and Any Place
Abstract / Description: 

In this presentation, I would like to discuss with you how to establish a sustainable and smart society through the internet of energy for connecting at any time and any place. I suspect that you have heard the phrase, "Internet of Energy" less often. The meaning of this phrase is simple. Because of a ubiquitous energy transmission system, you do not need to worry about a shortage of electric power. One of the most important items for establishing a sustainable society is [...]


"Inaugural Linvill Distinguished Seminar on Electronic Systems Technology," EE News, July 2018

 

Date and Time: 
Monday, January 14, 2019 - 4:30pm
Venue: 
Hewlett 200

Claude E. Shannon's 100th Birthday

Topic: 
Centennial year of the 'Father of the Information Age'
Abstract / Description: 

From UCLA Shannon Centennial Celebration website:

Claude Shannon was an American mathematician, electrical engineer, and cryptographer known as "the father of information theory". Shannon founded information theory and is perhaps equally well known for founding both digital computer and digital circuit design theory. Shannon also laid the foundations of cryptography and did basic work on code breaking and secure telecommunications.

 

Events taking place around the world are listed at IEEE Information Theory Society.

Date and Time: 
Saturday, April 30, 2016 - 12:00pm to Sunday, May 1, 2016 - 11:55am
Venue: 
N/A

NetSeminar

Topic: 
BlindBox: Deep Packet Inspection over Encrypted Traffic
Abstract / Description: 

SIGCOMM 2015, Joint work with: Justine Sherry, Chang Lan, and Sylvia Ratnasamy

Many network middleboxes perform deep packet inspection (DPI), a set of useful tasks which examine packet payloads. These tasks include intrusion detection (IDS), exfiltration detection, and parental filtering. However, a long-standing issue is that once packets are sent over HTTPS, middleboxes can no longer accomplish their tasks because the payloads are encrypted. Hence, one is faced with the choice of only one of two desirable properties: the functionality of middleboxes and the privacy of encryption.

We propose BlindBox, the first system that simultaneously provides both of these properties. The approach of BlindBox is to perform the deep-packet inspection directly on the encrypted traffic. BlindBox realizes this approach through a new protocol and new encryption schemes. We demonstrate that BlindBox enables applications such as IDS, exfiltration detection and parental filtering, and supports real rulesets from both open-source and industrial DPI systems. We implemented BlindBox and showed that it is practical for settings with long-lived HTTPS connections. Moreover, its core encryption scheme is 3-6 orders of magnitude faster than existing relevant cryptographic schemes.

Date and Time: 
Wednesday, November 11, 2015 - 12:15pm to 1:30pm
Venue: 
Packard 202

NetSeminar

Topic: 
Precise localization and high throughput backscatter using WiFi signals
Abstract / Description: 

Indoor localization holds great promise to enable applications like location-based advertising, indoor navigation, inventory monitoring and management. SpotFi is an accurate indoor localization system that can be deployed on commodity WiFi infrastructure. SpotFi only uses information that is already exposed by WiFi chips and does not require any hardware or firmware changes, yet achieves the same accuracy as state-of-the-art localization systems.

We then talk about BackFi, a novel communication system that enables high throughput, long range communication between very low power backscatter IoT sensors and WiFi APs using ambient WiFi transmissions as the excitation signal. We show via prototypes and experiments that it is possible to achieve communication rates of up to 5 Mbps at a range of 1 m and 1 Mbps at a range of 5 meters. Such performance is an order to three orders of magnitude better than the best known prior WiFi backscatter system.

Date and Time: 
Thursday, October 15, 2015 - 12:15pm to 1:30pm
Venue: 
Gates 104

NetSeminar

Topic: 
BlindBox: Deep Packet Inspection over Encrypted Traffic
Abstract / Description: 

SIGCOMM 2015, Joint work with: Justine Sherry, Chang Lan, and Sylvia Ratnasamy

Many network middleboxes perform deep packet inspection (DPI), a set of useful tasks which examine packet payloads. These tasks include intrusion detection (IDS), exfiltration detection, and parental filtering. However, a long-standing issue is that once packets are sent over HTTPS, middleboxes can no longer accomplish their tasks because the payloads are encrypted. Hence, one is faced with the choice of only one of two desirable properties: the functionality of middleboxes and the privacy of encryption.

We propose BlindBox, the first system that simultaneously provides both of these properties. The approach of BlindBox is to perform the deep-packet inspection directly on the encrypted traffic. BlindBox realizes this approach through a new protocol and new encryption schemes. We demonstrate that BlindBox enables applications such as IDS, exfiltration detection and parental filtering, and supports real rulesets from both open-source and industrial DPI systems. We implemented BlindBox and showed that it is practical for settings with long-lived HTTPS connections. Moreover, its core encryption scheme is 3-6 orders of magnitude faster than existing relevant cryptographic schemes.

Date and Time: 
Wednesday, October 7, 2015 - 12:15pm to 1:30pm
Venue: 
AllenX Auditorium

Pages

Statistics and Probability Seminars

Statistics Seminar: How much data is sufficient to learn high-performing algorithms?

Topic: 
How much data is sufficient to learn high-performing algorithms?
Abstract / Description: 

Algorithms often have tunable parameters that have a considerable impact on their runtime and solution quality. A growing body of research has demonstrated that data-driven algorithm design can lead to significant gains in runtime and solution quality. Data-driven algorithm design uses a training set of problem instances sampled from an unknown, application-specific distribution and returns a parameter setting with strong average performance on the training set. We provide a broadly applicable theory for deriving generalization guarantees for data-driven algorithm design, which bound the difference between the algorithm's expected performance and its average performance over the training set.

The challenge is that for many combinatorial algorithms, performance is a volatile function of the parameters: slightly perturbing the parameters can cause a cascade of changes in the algorithm’s behavior. Prior research has proved generalization bounds by employing case-by-case analyses of parameterized greedy algorithms, clustering algorithms, integer programming algorithms, and selling mechanisms. We uncover a unifying structure which we use to prove very general guarantees, yet we recover the bounds from prior research. Our guarantees apply whenever an algorithm's performance is a piecewise-constant, -linear, or — more generally — piecewise-structured function of its parameters. As we demonstrate, our theory also implies novel bounds for dynamic programming algorithms used in computational biology and voting mechanisms from economics.

This talk is based on joint work with Nina Balcan, Dan DeBlasio, Travis Dick, Carl Kingsford, and Tuomas Sandholm.

Date and Time: 
Tuesday, July 20, 2021 - 4:30pm

Statistics Department Seminar: High-dimensional and nonparametric estimation under 'local' information constraints

Topic: 
High-dimensional and nonparametric estimation under 'local' information constraints
Abstract / Description: 

Details unavailable at time of publishing. Please check Statistics pages for updated information.

Date and Time: 
Tuesday, June 29, 2021 - 4:30pm

Statistics Department Seminar: Balancing covariates in randomized experiments with the Gram-Schmidt Walk design

Topic: 
Balancing covariates in randomized experiments with the Gram-Schmidt Walk design
Abstract / Description: 

Details unavailable at time of publishing. Please check Statistics pages for updated information.

Date and Time: 
Tuesday, June 15, 2021 - 4:30pm

Statistics Department Seminar: Provable guarantees for self-supervised deep learning with spectral contrastive loss

Topic: 
Provable guarantees for self-supervised deep learning with spectral contrastive loss
Abstract / Description: 

Details unavailable at time of publishing. Please check Statistics pages for updated information.

Date and Time: 
Tuesday, June 1, 2021 - 4:30pm

Pages

SystemX

CANCELLED - SystemX Alliance presents "Autonomous Navigation of Miniature Distributed Space Systems"

Topic: 
Autonomous Navigation of Miniature Distributed Space Systems
Abstract / Description: 

Two key trends are revolutionizing the way humans conduct spaceflight, namely, the miniaturization of satellites (e.g., micro- and nano-satellites) and the distribution of payload tasks among multiple coordinated units (e.g., formation-flying, on-orbit servicing, fractionation, swarms). The combination of these approaches promises breakthroughs in space science (e.g., imaging of earth-like planets, characterization of gravitational waves), remote sensing (e.g., synthetic aperture radar interferometry, aeronomy, gravimetry), and space exploration (e.g., lifetime extension, assembly of structures, space debris removal). Irrespective of the specific application, future miniature distributed space missions require a high level of autonomy to maintain and reconfigure the relative motion of the participating vehicles within the prescribed accuracy and range of operations. Especially on small spacecraft, these requirements are hard to meet due to the limited resources, and the chief goal of current research and development is to pave the way for the autonomous Guidance, Navigation, & Control (GN&C) of "self-driving nanosatellites."

Leveraging the presenter's experience and contributions to recent satellite formation-flying and rendezvous missions in low earth orbit (TanDEM-X, Prisma, Biros), this presentation addresses the navigation algorithms under developments to enable a new class of miniature distributed space instruments (Starling, Visors, Swarm-Ex). The focus will be on autonomous vision-based and radio-frequency spaceborne absolute and relative navigation systems, including their training and validation using high-fidelity hardware-in-the-loop simulations and robotic testbeds at the Stanford's Space Rendezvous Laboratory.

Date and Time: 
Thursday, June 3, 2021 - 4:30pm

SystemX alliance presents "Mixed-Signal Circuits for tinyML Systems"

Topic: 
Mixed-Signal Circuits for tinyML Systems
Abstract / Description: 

Over the past decade, machine learning algorithms have been deployed in many cloud-centric applications. However, as the application space continues to grow, various algorithms are now embedded "closer to the sensor" and in wearable devices, eliminating the latency, privacy and energy penalties associated with cloud access. In this talk, I will review mixed-signal circuit techniques that can improve the energy efficiency of moderate-complexity, low-power machine learning inference algorithms. Specific examples include analog feature extraction for image and audio processing, as well as mixed-signal compute circuits for convolutional neural networks.

Date and Time: 
Thursday, May 27, 2021 - 4:30pm

SystemX Bonus Seminar: Industry 4.0: Making sense of signals

Topic: 
Industry 4.0: Making sense of signals
Abstract / Description: 

Industry 4.0 and smart manufacturing will be a critical driver of digital transformation. This inter-disciplinary domain can bring significant quality, productivity and cost benefits across sectors. Today's manufacturing environment is replete with hundreds of signals with petabytes of data. To get the highest ROI from Industry 4.0, the critical success-factor is to make sense of these signals. This seminar will focus on benefits and challenges of system-level ultra-low power sensing, ingesting high-frequency signals, as well as high SNR inferencing with AI/ML.

Date and Time: 
Tuesday, May 18, 2021 - 4:30pm

SystemX Alliance presents "Pervasive computing with everyday devices to build & sustain stress, resilience, and wellbeing"

Topic: 
Pervasive computing with everyday devices to build & sustain stress, resilience, and wellbeing
Abstract / Description: 

As society progresses towards increasing pervasive computing levels, I design and build technology-enabled solutions to repurpose everyday devices to help people build resilience and grow wellbeing. I leverage biological and behavioral knowledge to design systems that balance user needs and health outcomes while mitigating surveillance and agency risks. In this talk, I present my research on efficacious and engaging sensors and interventions necessary in the population and public health domains. I share a series of research projects exploring and validating novel ideas on passive sensors - less dependent on subjective surveys or wearables - and subtle interventions that minimize workflow disruption. I show the promise of repurposing existing signals from computing peripherals (i.e., mouse and trackpad) or cars (steering wheel) into "sensorless" sensors and repurposing existing media as just-in-time micro-interventions that can work across multiple scenarios and populations. I discuss how these data could be used in collaboration with domain experts to study topics as varied as the interaction between stress and productivity in office workers, burnout prevention among clinical practitioners, or the prevention of depression among rural health workers. Finally, grounded in theories from neuroscience and behavioral economics, I propose the evolution of everyday "mundane" devices, such as chairs, desks, cars, or even urban lights, into adaptive and autonomous wellbeing-optimizing interventions. I close with a discussion of the research needed to systematically study ethics in pervasive technology for resilience and wellbeing.

Date and Time: 
Thursday, May 20, 2021 - 4:30pm

SystemX presents "Quantum measurements for bio-imaging and precision sensing"

Topic: 
Quantum measurements for bio-imaging and precision sensing
Abstract / Description: 

Innovations in quantum science and technology have enabled previously unforeseen levels of measurement and control over simple quantum systems comprised of atoms, electrons and photons. This talk will illustrate some of these advances with examples relevant to bio-imaging and precision sensing. For example, so-called quantum non-demolition measurements provide a foundation for next generation electron microscopes capable of low-damage imaging of biological samples such as proteins. Related non-demolition protocols for atoms in high finesse optical cavities are used to produce entangled states of more than 1000 ultra-cold atoms. These states are exploited to improve the performance of atomic clocks. Finally, atom de Broglie wave interference experiments with atomic wavepackets separated by distances as large as 54 cm (image below is the resulting interference pattern) are being used for new tests of gravity and quantum mechanics. These examples, and their possible impact on basic science and technology, will be discussed.

Date and Time: 
Thursday, May 13, 2021 - 4:30pm

SystemX Alliance 2021 (Virtual) Spring Workshop

Topic: 
Future Directions of Semiconductor Technology - from Atoms to Applications
Abstract / Description: 

For this workshop, we are gathering visionaries and leaders of industry and academia to give us their perspectives followed by an open discussion. The audience will be Stanford faculty and students, and members of the Stanford SystemX Alliance. All talks will be by-invitation only.

Background: Because there is a shift from 2D scaling to a plethora of new materials and new devices for further performance gains, plus a concomitant shift from system-on-chip (SoC) to heterogeneous integration, chiplet partitioning, and advanced packaging, the roadmap ahead for the semiconductor industry is not as straightforward as it was in the past. Yet, there is tremendous value to industry and academia to know what the future of semiconductor technology will look like, and use that knowledge to plan for the future. The goal of this workshop is to have a forum for sharing perspectives and viewpoints on what future may hold in their respective domains. It is hoped that everyone in the food chain will gain from understanding the opportunities and challenges of an abstraction layer above and an abstraction layer below.

Date and Time: 
Wednesday, June 2, 2021 - 9:00am to Thursday, June 3, 2021 - 11:55am
Thursday, June 3, 2021 - 9:00am to Friday, June 4, 2021 - 11:55am
Monday, June 7, 2021 - 4:00pm to Tuesday, June 8, 2021 - 6:55pm

SystemX presents "“Tiny-but-tough” Gallium Nitride Sensors for Extreme Harsh Environments"

Topic: 
“Tiny-but-tough” Gallium Nitride Sensors for Extreme Harsh Environments
Abstract / Description: 

Gallium nitride (GaN) nanoelectronics have operated at temperatures as high as 1000°C making it a viable platform for robust space-grade ("tiny-but-tough") sensorsand electronics. In addition, there has been a tremendous amount of research and industrial investment in GaN as it is positioned to replace silicon in the billion-dollar (USD) power electronics industry, as well as the post-Moore microelectronics universe. Furthermore, the 2014 Nobel Prize in physics was awarded for pioneering research in GaN that led to the realization of the energy-efficient blue light-emitting diode (LED). Even with these major technological breakthroughs, we have just begun the "GaN revolution." New communities are adopting this nanoelectronic platform for a multitude of emerging device applications including the following: sensing, energy harvesting, actuation, and communication. In this talk, we will review and discuss the benefits of GaN's two-dimensional electron gas (2DEG) over silicon's p-n junction for space exploration applications (e.g., radiation-hardened, temperature-tolerant Venus instrumentation). In addition, we will discuss recent results that advance this nanoelectronic device platform for extreme-environment Internet-of-things (IoT)sensors for combustion and down-hole monitoring.

Date and Time: 
Thursday, May 6, 2021 - 4:30pm

SystemX presents "Use of the Roadmap (IRDS) Method to Overcome Industry Inflection Points"

Topic: 
Use of the Roadmap (IRDS) Method to Overcome Industry Inflection Points
Abstract / Description: 

In 1965 Gordon Moore, co-founder of Fairchild Semiconductors and also Intel Corporation, predicted that the number of transistors per die would double every year until 1975 if specific actions were taken to eliminate any inflection points. This represents the first example of a long-term roadmap for semiconductors.

By 1975 data showed his predictions to be correct. He then predicted that the number of transistors would double every two years for the foreseeable future if appropriate actions were taken. Following Gordon Moore's demonstrated success of the roadmap methodology, the National Technology Roadmap for Semiconductors (NTRS) was formed in 1991 in the US; the roadmap became international (ITRS) in 1998 to collectively address a major inflection point (i.e., the end of the planar MOS silicon gate process). Following the ITRS recommendations the semiconductor industry sequentially developed in a very timely fashion Strained Silicon, High-K/Metal-Gate and FinFET into manufacturing. Beginning with 2016 the scope of the roadmap was extended to also include System and Architecture as well as trends on Cryogenic Electronics and Quantum Information Processing. In the not-too-distant future 2D space in ICs will reach fundamental limits and fully utilization of the third dimension (3D) to increase "functionality density" has already started. Furthermore homogeneous and heterogeneous integration of multiple technologies is already in progress and will become the way of the future. New recommendations for actions to be taken in the next 15 years have been formulated and will be presented.

Date and Time: 
Thursday, April 29, 2021 - 4:30pm

SystemX presents "How to Think About AI Ethics: A New, Systemic Framework"

Topic: 
How to Think About AI Ethics: A New, Systemic Framework
Abstract / Description: 

Ethics is often messy and chaotic—it's hard to see the method in the madness. In contrast to the way ethics is usually taught, this talk will offer a new framework that lays out the process or ethics workflow, in order to systematically think through the issues in technology ethics, especially AI ethics. We will also look at a few AI case studies in applying this framework.

Date and Time: 
Thursday, April 22, 2021 - 4:30pm

SystemX presents "How to Compute with Schrödinger's Cat: An Introduction to Quantum Computing"

Topic: 
How to Compute with Schrödinger's Cat: An Introduction to Quantum Computing
Abstract / Description: 

The success of the abstract model of computation, in terms of bits, logical operations, algorithms, and programming language constructs makes it easy to forget that computation is a physical process. Our cherished notions of computation and information are grounded in classical mechanics, but the physics of our universe is quantum. A natural question to ask is how computation would change if we adopted a quantum mechanical, instead of a classical mechanical, model of computation. In the early 80s, Richard Feynman, Yuri Manin, and others recognized that certain quantum effect could not be simulated efficiently on conventional computers. This observation led researchers to speculate that perhaps such quantum effect could be used to speed up computation more generally. Slowly, a new picture of computation arose, one that gave rise to a variety of faster algorithms, novel cryptographic mechanisms, and alternative methods of communication. In the first part of the talk, I will introduce key concepts underlying quantum computing and correct misconceptions. In the second part of the talk, I will discuss research being done by NASA's QuAIL group on quantum algorithms, quantum supremacy, elucidating quantum resources for computation, quantum programming and compilation, quantum-inspired classical algorithms, and assessing future applications of quantum computing, all within the broader context of the rapidly evolving field.

Date and Time: 
Thursday, April 15, 2021 - 4:30pm

Pages

Subscribe to RSS - Seminar / Colloquium