EE Student Information

The Department of Electrical Engineering supports Black Lives Matter. Read more.

• • • • •

EE Student Information, Spring Quarter through Academic Year 2020-2021: FAQs and Updated EE Course List.

Updates will be posted on this page, as well as emailed to the EE student mail list.

Please see Stanford University Health Alerts for course and travel updates.

As always, use your best judgement and consider your own and others' well-being at all times.

Graduate

Fuse!

Topic: 
volleyball
Abstract / Description: 

Happy Spring Quarter!! The weather is beautiful, so let's take advantage of it and play some volleyball! We'll set up volleyball nets on the grass outside of Packard—if you want to help set up, come at 3:45, otherwise you can join us at 4 to play! If volleyball isn't your thing, you can still come and enjoy the sunshine with us!

Hopefully we'll see you there!

-------------

More about Fuse: ee.stanford.edu/student-resources/student-organizations

Date and Time: 
Friday, April 17, 2015 - 4:00pm to 6:30pm
Venue: 
Packard lawn

GSEE (Graduate Students in Electrical Engineering) Event

Topic: 
Meditation Session
Abstract / Description: 

GSEE is hosting a free 1-hour meditation session. Please register; there are a limited number of spots available. First-come first-serve and affiliation with the EE department. If you are interested, please sign up here (expired). We will let you know by next Wednesday, 4/22/15.

Date and Time: 
Friday, April 24, 2015 - 4:00pm to 5:00pm
Venue: 
TBA

IT-Forum

Topic: 
Sub-optimality of the Han-Kobayashi region for the interference channel: Remarks on computing regions expressed by auxiliaries
Abstract / Description: 

A computable characterization of the capacity region for the two receiver interference channel, along with that of the two receiver broadcast channel, constitute two of the most fundamental open questions in network information theory. A computable achievable region proposed by Han and Kobayashi (1981) was the best-known achievable region and its optimality or sub-optimality had not been established prior to this work. In this work, we show that coding across time-slots can improve on Han-Kobayashi region (also known as multi-letter inner bounds).

The talk will be mainly about the various ideas and developments that eventually led to this result. It touches upon various computational aspects of computable regions or in other words, the study of extremal auxiliaries.

A similar result about Marton's achievable region for broadcast channels has not been found yet; and the difficulty lies precisely in the narrowing of extremal auxliaries, although significant progress has been made in this direction in the past few years.

This will be a whiteboard talk and though there is a rough outline, the talk can be steered by the audience.

Date and Time: 
Friday, May 15, 2015 - 1:00pm to 2:00pm
Venue: 
Packard 202

IT-Forum: Social Learning in Decision-Making Groups

Topic: 
Social Learning in Decision-Making Groups
Abstract / Description: 

People have always been influenced by the opinions of their acquaintances. Increasingly, through recommendations and ratings provided on all sorts of goods and services, people are also influenced by the opinions of people that are not even acquaintances. This ubiquity of the sharing of opinions has intensified the interest is the concept of herding (or informational cascades) introduced in 1992. While agents in most previous works have only individualistic goals, this talk focuses on social influence among agents in two collaborative settings.

We consider agents that perform Bayesian binary hypothesis testing and, in addition to their private signals, observe the decisions of earlier-acting agents. In the first setting, each decision has its own corresponding Bayes risk. Each agent affects the minimum possible Bayes risk for subsequent agents, so an agent may have a mixed objective including her own Bayes risk and the Bayes risks of subsequent agents; we demonstrate her tension between being informative to other agents and being right in her own decisions, and we show that she is more informative to others when she is open minded. In the second setting, opinions are aggregated by voting, and all agents aim to minimize the Bayes risk of the team's decision. We show that social learning is futile when the agents observe conditionally independent and identically distributed private signals (but not merely conditionally independent private signals) or when the agents require unanimity to make a decision. Our experiments with human subjects suggest that when opinions of people with equal qualities of information are aggregated by voting, the ballots should be secret. They have also raised questions about rationality and trust.

Date and Time: 
Friday, April 24, 2015 - 1:00pm to 2:00pm
Venue: 
Packard 202

IT-Forum: Distributed Estimation of Generalized Matrix Rank: Efficient Algorithms and Lower Bounds

Topic: 
Distributed Estimation of Generalized Matrix Rank: Efficient Algorithms and Lower Bounds
Abstract / Description: 

We study the following generalized matrix rank estimation problem: given an n×n matrix and a constant c≥0, estimate the number of eigenvalues that are greater than c. In the distributed setting, the matrix of interest is the sum of m matrices held by separate machines. We show that any deterministic algorithm solving this problem must communicate Ω(n^2) bits, which is order-equivalent to transmitting the whole matrix. In contrast, we propose a randomized algorithm that communicates only O(n) bits. The upper bound is matched by an Ω(n) lower bound on the randomized communication complexity. We demonstrate the practical effectiveness of the proposed algorithm with some numerical experiments.

Date and Time: 
Friday, April 17, 2015 - 1:00pm to 2:00pm
Venue: 
Packard 202

Applied Physics/Physics Colloquium: Special Seminar

Topic: 
The Phases of Hard Sphere Systems
Abstract / Description: 

One of the simplest models of statistical mechanics is a system of identical hard sphere particles, placed inside a box whose energy and volume are free to fluctuate in response to an environment characterized by a temperature and pressure. A dimensionless pressure parameter p is formed from a combination of the sphere radius, the temperature, and the pressure of the environment. Qualitatively different average properties of the sphere packing are controlled by this dimensionless pressure: a "disordered" gas phase at small p and an "ordered" crystalline phase at large p. Since all (nonintersecting) sphere configurations are isoenergetic, the mechanism for crystalline ordering, called "order by disorder", is purely entropic in nature. Sphere packings having the highest possible density correspond to the limit of infinite p.


In addition to providing a physicist's intuition on the existence of phases in the hard sphere system, this tutorial talk will also touch on topics of mathematical interest. In the limit of many spheres (so "magic number" effects are minimized), does statistical mechanics distinguish among the different densest structures that arise in three dimensions? In higher dimensions, where much less is known, might we expect more than two phases, or perhaps just a single phase?

Applied Physics event page

Date and Time: 
Wednesday, April 22, 2015 - 4:00pm to 5:00pm
Venue: 
Building 380, Room 384-I

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

Pages

Subscribe to RSS - Graduate