Graduate

IEEE SCV-WIE presents Career Panel Talk: Advising My Younger Self

Topic: 
Career Panel Talk: Advising My Younger Self
Abstract / Description: 

The event will start with Career Perspectives from women who hold varied experiences both in industry and academia, facilitated by an IEEE-SCV-WIE Board Member and continue with the Q&A. The focus of the panel discussion will be around how to begin and maintain a passionate career track

Please register

  • Moderator: Karen Horovitz, IEEE-SCV-WIE Board
  • Panelists: Kitty Yeung, Liliane Peters, Mayuri Kulkarni, Patsy Price, and Sneha Prasad
Date and Time: 
Friday, May 25, 2018 - 4:00pm
Venue: 
Packard 202

SmartGrid Seminar: Trends in Electric Power Distribution System Analysis at PNNL

Topic: 
Trends in Electric Power Distribution System Analysis at PNNL
Abstract / Description: 

Pacific Northwest National Laboratory (PNNL) originated and continues to maintain one of the two leading open-source distribution system simulators, called GridLAB-D, which has been downloaded 80,000+ times world-wide. While it continues to improve core functionality, PNNL is placing more emphasis recently on GridLAB-D as part of a development platform, improving its interoperability and opening the software up to more customization by researchers. This talk will cover two ongoing open-source development projects, funded by the U. S. Department of Energy, that incorporate and extend GridLAB-D. One of these projects is also expected to contribute distribution feeder model conversion tools for a new California Energy Commission project headed by SLAC. Highlights of the talk will include:

  • Transactive energy simulation platform, at tesp.readthedocs.io/en/latest
  • GridAPPS-D application development platform, at gridappsd.readthedocs.io/en/latest
  • Evole GridLAB-D's co-simulation support from FNCS interface, to a multi-lab interface called HELICS compliant with Functional Mockup Interface (FMI): https://github.com/GMLC-TDC/HELICS-src
  • Leveraging new capabilities for large-building simulation in JModelica, power flow analysis in OpenDSS, and transactive energy system agents in Python
  • Implementation and use of the Common Information Model (CIM) in a NoSQL triple-store database for standardized feeder model conversion
  • Comparison of different types of stochastic modeling for load and distributed energy resource (DER) output variability, and its impact on feeder model order reduction and state estimation
  • Special system protection example concerns on urban secondary networks with high penetration of DER
Date and Time: 
Thursday, May 24, 2018 - 1:30pm
Venue: 
Y2E2 111

Fully Autonomous Vehicles: Removing the Front-Seat Driver

Topic: 
Fully Autonomous Vehicles: Removing the Front-Seat Driver
Abstract / Description: 

The automotive industry is bracing for a new surge of progress and opportunity, with billions of dollars on the line. The potential is huge. People have been dreaming about owning a self-driving car for decades, and recent sales have shown that consumers are willing to spend big bucks for this dream.

The final push to widespread availability of fully autonomous vehicles has begun. Testing has moved onto public roads, and in the next few years, several vendors plan to release autonomous vehicles in limited-use "Level 4" situations (taxis, transit companies, fleet operations, freeway-only driving). As the technology advances from "assisted driving" to true self-driving, consumer-ready vehicles for general road travel (Level 5) will be tailgating right behind.

In this panel, industry-leading startups, entrepreneurs, and investors will explore the technologies and challenges related to realizing the dream of a fully autonomous self-driving vehicle — i.e. a vehicle that a human does not need to monitor and that can completely take over all driving responsibilities.

Panelists:
- Sterling Anderson, Co-Founder and Chief Product Officer, Aurora
Named "America's Hottest Self-Driving Startup" by Wired
- Alexei Andreev, Managing Director, Autotech Ventures
Focused on the $3T ground transportation sector
- Ivan Mihov, Director of Program Management, Zoox
Valued at over $1B by Crunchbase
- Rob Coneybeer, Managing Director, Shasta Ventures (Moderator)

FREE for Stanford ID holders. Registration required by 5/23. Please email  briana.burrows@stanford.edu  for a ticket.

Date and Time: 
Thursday, May 24, 2018 - 6:00pm
Venue: 
Hewlett 200

Social Entrepreneurship and Tech Innovation Advancing Sustainable Development Goals in South Asia

Topic: 
Social Entrepreneurship and Tech Innovation Advancing Sustainable Development Goals in South Asia
Abstract / Description: 

Panelists:
- Rikin Gandhi, Co-founder/Executive Director of Digital Green
- Radhika Shah, Co-President of Stanford Angels & Entrepreneurs
- Dr. Richard Dasher, Director of the US-Asia Technology Management Center (Moderator)

Rikin Gandhi is co-founder and executive director of Digital Green, a global development organization that empowers smallholder farmers to lift themselves out of poverty by harnessing the collective power of technology and grassroots-level partnerships. He began his career at Oracle, where he received patents for linguistic search algorithms that he helped develop. Later, he joined Microsoft Research India's Technology for Emerging Markets team, where he researched ways to amplify the effectiveness of agricultural development globally. While traveling around India's rural communities, Gandhi developed a passion for helping the country's rural farmers. That passion then became his career: in 2006, he co-founded what is now Digital Green. Gandhi holds a master's in aeronautical and astronautical space engineering from Massachusetts Institute of Technology, and a bachelor's in computer science from Carnegie Mellon University.

Radhika Shah is Co-President of Stanford Angels & Entrepreneurs. She is an angel/impact tech investor also passionate about civic engagement, community building & transformative social change. She is an advisor to the Sustainable Development Goals Philanthropy Platform and Founding Chair of the Tech Advisory Group for Stanford Handa Center for Human Rights. She also sits on the International Advisory Network for Business and Human Rights Resource Centre. She is passionate about women's rights, education, social justice, environment since her childhood, growing up under the influence of the Gandhi Ashram in India. Radhika co-founded former Ashoka SV chapter, is on SV Leadership Council for Action for India, an advisor/mentor to CGNetSwara.org, Samasource.org.

See https://asia.stanford.edu/social-entrepreneurship-may-22-2018/ for details.

Date and Time: 
Tuesday, May 22, 2018 - 4:30pm
Venue: 
Skilling Auditorium, 494 Lomita Mall

ISL Colloquium: Finite Sample Guarantees for Control of an Unknown Linear Dynamical System

Topic: 
Finite Sample Guarantees for Control of an Unknown Linear Dynamical System
Abstract / Description: 

In principle, control of a physical system is accomplished by first deriving a faithful model of the underlying dynamics from first principles, and then solving an optimal control problem with the modeled dynamics. In practice, the system may be too complex to precisely characterize, and an appealing alternative is to instead collect trajectories of the system and fit a model of the dynamics from the data. How many samples are needed for this to work? How sub-optimal is the resulting controller?

In this talk, I will shed light on these questions when the underlying dynamical system is linear and the control objective is quadratic, a classic optimal control problem known as the Linear Quadratic Regulator. Despite the simplicity of linear dynamical systems, deriving finite-time guarantees for both system identification and controller performance is non-trivial. I will first talk about our results in the "one-shot" setting, where measurements are collected offline, a model is estimated from the data, and a controller is synthesized using the estimated model with confidence bounds. Then, I will discuss our recent work on guarantees in the online regret setting, where noise injected into the system for learning the dynamics needs to trade-off with state regulation.

This talk is based on joint work with Sarah Dean, Horia Mania, Nikolai Matni, and Benjamin Recht.

Date and Time: 
Thursday, May 24, 2018 - 4:15pm
Venue: 
Packard 101

Applied Physics/Physics Colloquium: The IceCube Neutrino Observatory and the Beginning of Neutrino Astrophysics

Topic: 
The IceCube Neutrino Observatory and the Beginning of Neutrino Astrophysics
Abstract / Description: 

The IceCube Neutrino Observatory is the world's largest neutrino detector, instrumenting a cubic kilometer of ice at the geographic South Pole. IceCube was designed to detect high-energy astrophysical neutrinos from potential cosmic ray acceleration sites such as active galactic nuclei, gamma ray bursts and supernova remnants. IceCube announced the detection of a diffuse flux of astrophysical neutrinos in 2013, including the highest energy neutrinos ever detected. The sources of these neutrinos are as yet unknown, and IceCube continues to collect data and to collaborate with multi messenger partners in order to explore the neutrino sky. I will discuss the latest results from IceCube and discuss prospects for future upgrades and expansions of the detector.

Date and Time: 
Tuesday, May 22, 2018 - 4:30pm
Venue: 
Hewlett 201

IT Forum: Phase transitions in generalized linear models

Topic: 
Phase transitions in generalized linear models
Abstract / Description: 

This is joint work with Jean Barbier, Florent Krzakala, Nicolas Macris and Lenka Zdeborova.

We consider generalized linear models (GLMs) where an unknown $n$-dimensional signal vector is observed through the application of a random matrix and a non-linear (possibly probabilistic) componentwise output function.

We study the models in the high-dimensional limit, where the observations consists of $m$ points, and $m/n \to \alpha > 0$ as $n \to \infty$. This situation is ubiquitous in applications ranging from supervised machine learning to signal processing.

We will analyze the model-case when the observation matrix has i.i.d. elements and the components of the ground-truth signal are taken independently from some known distribution.

We will compute the limit of the mutual information between the signal and the observations in the large system limit. This quantity is particularly interesting because it is related to the free energy (i.e. the logarithm of the partition function) of the posterior distribution of the signal given the observations. Therefore, the study of the asymptotic mutual information allows to deduce the limit of important quantities such as the minimum mean squared error for the estimation of the signal.

We will observe some phase transition phenomena. Depending on the noise level, the distribution of the signal and the non-linear function of the GLM we may encounter various scenarios where it may be impossible / hard (only with exponential-time algorithms) / easy (with polynomial-time algorithms) to recover the signal.

Date and Time: 
Friday, May 18, 2018 - 1:15pm
Venue: 
Packard 202

ISL Colloquium & IT Forum: Random initialization and implicit regularization in nonconvex statistical estimation

Topic: 
Random initialization and implicit regularization in nonconvex statistical estimation
Abstract / Description: 

Recent years have seen a flurry of activities in designing provably efficient nonconvex procedures for solving statistical estimation / learning problems. Due to the highly nonconvex nature of the empirical loss, state-of-the-art procedures often require suitable initialization and proper regularization (e.g. trimming, regularized cost, projection) in order to guarantee fast convergence. For vanilla procedures such as gradient descent, however, prior theory is often either far from optimal or completely lacks theoretical guarantees.

This talk is concerned with a striking phenomenon arising in two nonconvex problems (i.e. phase retrieval and matrix completion): even in the absence of careful initialization, proper saddle escaping, and/or explicit regularization, gradient descent converges to the optimal solution within a logarithmic number of iterations, thus achieving near-optimal statistical and computational guarantees at once. All of this is achieved by exploiting the statistical models in analyzing optimization algorithms, via a leave-one-out approach that enables the decoupling of certain statistical dependency between the gradient descent iterates and the data. As a byproduct, for noisy matrix completion, we demonstrate that gradient descent achieves near-optimal entrywise error control.

Date and Time: 
Wednesday, May 23, 2018 - 4:15pm
Venue: 
Building 370

SystemX Seminar: Intracellular recording of thousands of connected neurons on a silicon chip

Topic: 
Intracellular recording of thousands of connected neurons on a silicon chip
Abstract / Description: 

Massively parallel, intracellular recording of a large number of neurons across a network is a great technological pursuit in neurobiology, but it has not been achieved. The intracellular recording by the patch clamp electrode boasts unparalleled sensitivity that can measure down to sub-threshold synaptic events, but it is too bulky to be implemented into a dense massive-scale array: so far only ~10 parallel patch recordings have been possible. Optical methods––e.g., voltage-sensitive dyes/proteins––have been developed in hopes of parallelizing intracellular recording, but they have not been able to perform recording from more than ~30 neurons in parallel. As an opposite example, the microelectrode array can record from many more neurons, but this extracellular technique has too low a sensitivity to tap into synaptic activities. In this talk, I would like to share our on-going effort, a silicon chip that conducts intracellular recording from thousands of connected mammalian neurons in vitro, and discuss applications in high-throughput screening, functional connectome mapping, neuromorphic engineering, and data science.

Date and Time: 
Tuesday, May 15, 2018 - 2:00pm
Venue: 
Allen 101X

IT-Forum: Tight sample complexity bounds via dualizing LeCam's method

Topic: 
Tight sample complexity bounds via dualizing LeCam's method
Abstract / Description: 

In this talk we consider a general question of estimating linear functional of the distribution based on the noisy samples from it. We discover that the (two-point) LeCam lower bound is in fact achievable by optimizing bias-variance tradeoff of an empirical-mean type of estimator. We extend the method to certain symmetric functionals of high-dimensional parametric models.

Next, we apply this general framework to two problems: population recovery and predicting the number of unseen species. In population recovery, the goal is to estimate an unknown high-dimensional distribution (in $L_\infty$-distance) from noisy samples. In the case of \textit{erasure} noise, i.e. when each coordinate is erased with probability $\epsilon$, we discover a curious phase transition in sample complexity at $\epsilon=1/2$. In the second (classical) problem, we observe $n$ iid samples from an unknown distribution on a countable alphabet and the goal is to predict the number of new species that will be observed in the next (unseen) $tn$ samples. Again, we discover a phase transition at $t=1$. In both cases, the complete characterization of sample complexity relies on complex-analytic methods, such as Hadamard's three-lines theorem.

Joint work with Yihong Wu (Yale).

Date and Time: 
Friday, May 4, 2018 - 1:15pm
Venue: 
Packard 202

Pages

Subscribe to RSS - Graduate