Graduate

ISL Colloquium: Learning sparse polynomials and graphs using coding theory tools

Topic: 
Learning sparse polynomials and graphs using coding theory tools
Abstract / Description: 

Learning sparse polynomials from random samples is a notorious problem in learning theory. We show that
if the coefficients are in general position we can learn such functions efficiently. We show how this problem
has applications on learning the structure of unknown graphs by observing the values of graph cuts. We
further discuss our on-going work on learning variable interactions (quadratic polynomials) very efficiently.
Our techniques are coding theoretic and boil down to solving noisy linear equations over a finite field.

Based on joint work with Murat Kocaoglu, Karthik Shanmugam, and Adam Klivans.

Date and Time: 
Thursday, April 16, 2015 - 4:15pm to 5:15pm
Venue: 
Packard 101

Information Systems Lab Colloquium: Point-Map Probabilities of a Point Process

Topic: 
Point-Map Probabilities of a Point Process
Abstract / Description: 

A compatible point-shift f maps, in a translation invariant way, each point of a stationary point process Φ to some point of Φ. It is fully determined by its associated point-map, g^f, which gives the image of the origin by f. The initial question of this paper is whether there exist probability measures which are left invariant by the translation of −g^f. The point-map probabilities of Φ are defined from the action of the semigroup of point-map translations on the space of Palm probabilities, and more precisely from the compactification of the orbits of this semigroup action. If the point-map probability is uniquely defined, and if it satisfies certain continuity properties, it then provides a solution to the initial question. Point-map probabilities are shown to be a strict generalization of Palm probabilities: when the considered point-shift f is bijective, the point-map probability of Φ boils down to the Palm probability of Φ. When it is not bijective, there exist cases where the point-map probability of Φ is absolutely continuous with respect to its Palm probability, but there also exist cases where it is singular with respect to the latter. A criterium of existence of the point-map probabilities of a stationary point process is also provided. The results are illustrated by a few examples.

This is joint work with Mir-Omid Haji-Mirsadeghi, Sharif University.

ADDITIONAL DETAILS

Date and Time: 
Monday, April 13, 2015 - 4:15pm to 5:15pm
Venue: 
Sequoia Hall, Room 200

Information Systems Lab Colloquium: Information Relaxations and Duality in Stochastic Dynamic Programs

Topic: 
Information Relaxations and Duality in Stochastic Dynamic Programs
Abstract / Description: 

In this talk, we discuss the information relaxation approach for obtaining bounds on the performance of optimal policies in stochastic dynamic programs (DP). This approach involves relaxing the DP information structure and incorporating a penalty that punishes the use of additional information. We first provide an overview of some basic theory for the general approach. We then discuss how to apply the method in two broad classes of problems. For DPs with a convex structure, we show how to use gradient penalties and convex optimization to apply the method, and illustrate with an application in network revenue management. For infinite horizon MDPs, we show how to apply the method with change of measure techniques, and illustrate with an application in service allocation for a multiclass queue with convex delay costs. As we discuss, in both cases, the method provides tighter bounds than bounds from other relaxation methods, such as Lagrangian relaxations.

Date and Time: 
Wednesday, April 15, 2015 - 4:00pm to 5:00pm
Venue: 
Patterson Bldg, Room P107

EE380 Computer Systems Colloquium: Intel Software Guard Extensions Innovative Instructions for Next Generation Isolated Execution

Topic: 
Intel Software Guard Extensions Innovative Instructions for Next Generation Isolated Execution
Abstract / Description: 

This talk describes Intel's Software Guard Extensions (SGX) technology. SGX provides new tools and hardware facilities to software developers to protect an application's secrets. In today's computing environment the ability to keep a secret requires the integrity of millions of line of software in the OS, VMM, and application. SGX creates a trusted environment called an enclave inside the application. An enclave provides an ability to protect the secret without dependency on the integrity of any other code. The talk will describe the programming environment, instruction set, and hardware facilities which make up the SGX architecture.


ABOUT THE COLLOQUIUM:

See the Colloquium website, http://ee380.stanford.edu, for scheduled speakers, FAQ, and additional information. Stanford and SCPD students can enroll in EE380 for one unit of credit. Anyone is welcome to attend; talks are webcast live and archived for on-demand viewing over the web.

Date and Time: 
Wednesday, April 15, 2015 - 4:15pm to 5:15pm
Venue: 
Gates B1

Applied Physics/Physics Colloquium: Advanced LIGO

Topic: 
Advanced LIGO: The Coming Dawn of Gravitational Wave Physics and Astronomy
Abstract / Description: 

For more than 50 years, scientists have endeavored to detect gravitational waves from galactic and extra-galactic compact astrophysical sources such as supernovae and merging neutron stars and black holes. Beginning in the 1960's, a series of ever more sensitive instruments were constructed to search for them. Alas, no gravitational waves have yet been found.

With Advanced LIGO coming online in 2015, we have good reason to believe that is about to change. In this colloquium I'll discuss gravitational waves, what makes them so interesting and challenging to detect, and how the Advanced LIGO interferometers will hunt for them. Gravitational wave science cuts across a broad swath of disciplines: general relativity, classical and quantum optics, materials science, and multi-messenger astronomy, so there should be something in the talk for everyone.

Date and Time: 
Tuesday, May 26, 2015 - 4:00pm to 5:00pm
Venue: 
Hewlett 201

Applied Physics/Physics Colloquium

Topic: 
Topological Insulators beyond Band Theory
Abstract / Description: 

Our appreciation of insulating states of quantum matter has been deepened in the last decade by the theoretical prediction and subsequent experimental discovery of topological insulators. Much of the theoretical discussion of these phases is informed by free electron band theory. In this talk, I will describe recent progress in generalizing the concept of topological insulation to strongly interacting electronic systems. I focus on a minimal generalization known as symmetry-protected topological (SPT) phases. Just like the familiar topological insulators, these states have a bulk gap and no exotic excitations, but have nontrivial surface states that are protected by symmetry. I will show how interactions enable many SPT phases for spin-orbit coupled three dimensional electronic insulators that have no analog in band theory. I describe their physical properties and experimental fingerprints. More generally, studies of such SPT phases represent possibly the simplest context for the interplay between interactions, symmetry, and topology in states of quantum matter. I will describe how the insights obtained provide fresh viewpoints on a number of other frontier theoretical problems in quantum condensed matter physics.

Date and Time: 
Tuesday, May 19, 2015 - 4:00pm to 5:15pm
Venue: 
Hewlett 201

Applied Physics/Physics Colloquium

Topic: 
Searching for the First Stars
Abstract / Description: 

Through observations of the CMB, we have a clear picture of the conditions in the universe at the point where atoms first combined and the universe became transparent. From deep observations of galaxies, we can study the universe at a z of ~ 6. Here, we find galaxies, smaller and more irregular than in our local universe, but recognizable as galaxies. In the critical period between recombination and z~6, the first stars and galaxies were born. Their radiation reionized the universe, and their nucleosynthetic products contaminated the interstellar matter from which later generations of stars would form with heavy elements. I will review the investigations that have searched for the diffuse light from this epoch, starting with total power observations such as those done by DIRBE on COBE to more recent observations of spatial correlations by the GSFC, Akiri, Irvine, and Caltech groups. There has been rapid progress in this research. Spatial correlations between the near IR emission and Chandra soft X-ray emission suggests an origin in energetic systems, while large scale spatial correlations seen in the near IR suggest an origin in stars stripped from galaxies. The next years will see these conflicts resolved and may reveal a first clear view into the concealed early history of the universe.

Date and Time: 
Tuesday, May 12, 2015 - 4:00pm to 5:00pm
Venue: 
Hewlett 201

Applied Physics/Physics Colloquium

Topic: 
The Story of Single Molecules, from Early Spectroscopy in Solids to Super-Resolution Nanoscopy in Cells and Beyond
Abstract / Description: 

More than 25 years ago, low temperature experiments aimed at establishing the ultimate limits to optical storage in solids led to the first optical detection and spectroscopy of a single molecule in the condensed phase. At this unexplored ultimate limit, many surprises occurred where single molecules showed both spontaneous changes (blinking) and light-driven control of emission, properties that were also observed in 1997 at room temperature with single green fluorescent protein variants. In 2006, PALM and subsequent approaches showed that the optical diffraction limit of ~200 nm can be circumvented to achieve super-resolution fluorescence microscopy, or nanoscopy, with relatively nonperturbative visible light. Essential to this is the combination of single-molecule fluorescence imaging with active control of the emitting concentration and sequential localization of single fluorophores decorating a structure. Super-resolution microscopy has opened up a new frontier in which biological structures and behavior can be observed in live cells with resolutions down to 20-40 nm and below. Examples range from protein superstructures in bacteria to bands in actin filaments to details of the shapes of amyloid fibrils and much more. Current methods development research addresses ways to extract more information from each single molecule such as 3D position and orientation, in thick cells. Still, it is worth noting that in spite of all the focus on super-resolution, even in the "conventional" single-molecule tracking regime where the motions of individual biomolecules are recorded in solution or in cells rather than the shapes of extended structures, much can still be learned about biological processes.

Date and Time: 
Tuesday, April 28, 2015 - 4:00pm to 5:00pm
Venue: 
Hewlett 201

Applied Physics/Physics Colloquium: Quantum Entanglement in Higher Dimensions

Topic: 
Quantum Entanglement in Higher Dimensions
Abstract / Description: 

2015 Robert Hofstadter Memorial Lecturer

Many fundamental experiments in quantum information have been performed with qubits, i.e. in rather low dimensional Hilbert spaces. Various experimental techniques have recently opened up discrete higher dimensions for experiments. These are particularly multi-mode interference and photon states with more complex wavefronts, like orbital angular momentum (OAM) states.

I will present some recent results focusing on OAM states and on verifications of entanglement in very high dimensions. So far, entanglement of quantum states with quantum numbers around 1,000 and of superposition in more than 100-dimensional Hilbert space have been realized. This may shed interesting light on the question of the quantum-classical transition. I will also discuss recent experiments for using some of these states in quantum communication with higher alphabets.

Entangled quantum states also provide novel ways for nonlocal imaging. Most recently, it was seen that one can obtain imaging where the photon interacting with the object does not even have to be detected. This exploits the fact that the phase of a product state cannot be localized.

Date and Time: 
Tuesday, April 21, 2015 - 4:00pm to 5:00pm
Venue: 
Hewlett 201

Pages

Subscribe to RSS - Graduate