<< Monday, May 01, 2017 >>

Monday, May 1, 2017

Computational Challenges in Machine Learning

Workshop | May 1 – 5, 2017 every day |  Calvin Laboratory (Simons Institute for the Theory of Computing)

 Simons Institute for the Theory of Computing

The aim of this workshop will be to bring together a broad set of researchers looking at algorithmic questions that arise in machine learning. The primary target areas will be large-­scale learning and algorithms for Bayesian estimation.

Santosh Vempala (Georgia Institute of Technology; chair), David Blei (Columbia University), Katherine Heller (Duke University), John Langford...   More >

  Register online

Dissertation Talk: Provably Secure Computing using Certified Software and Trusted Hardware

Presentation | May 1 | 11 a.m.-12 p.m. | 531 Cory Hall

 Rohit Sinha

 Electrical Engineering and Computer Sciences (EECS)

Security-critical applications constantly face threats from exploits in lower computing layers 
such as the OS and Hypervisor, or even attacks from malicious datacenter insiders.
This dissertation explores building cloud applications with provable security guarantees, 
including only these hardware primitives (i.e. nearly zero software) in the trusted computing base. 

$K$-theory seminar: The Chern Character in twisted $K$-theory

Seminar | May 1 | 11:10 a.m.-12:30 p.m. | Evans Hall, 740 and then 959

 German Stefanich, UC Berkeley

 Department of Mathematics

We will recall the construction of the (equivariant) Chern character in $K$-theory and proceed to construct its twisted version.

Graduate Student Seminar

Seminar | May 1 | 12-1:30 p.m. | 489 Minor Hall

 Rachel Albert, PhD Candidate (O’Brein Lab); Paul Cullen (Flanagan Lab)

 Neuroscience Institute, Helen Wills

Rachel Albert's talk title is:

Latency Requirements for Foveated Rendering


Paul Cullen's talk title is:

Under Pressure: Understanding Glaucoma through the Lens of Astrocyte Reactivity

Anthropology on the Frontlines: Honoring the work of Nancy Scheper-Hughes

Conference/Symposium | May 1 | 1-7 p.m. |  Alumni House

 Marcelo Suarez-Orozco, Dean of Education, UCLA; Juan Thomas Ordóñez, Professor of Anthropology, Universidad del Rosario, Bogotá; Naomar Monteriro Almeida-Filho, Chancellor, Federal University Southern Bahia, Brazil; Sherry Ortner, Distinguished Professor of Anthropology, UCLA; Kimberly Theidon, Professor of Anthropology, Fletcher School of Diplomacy; Francisco J. Ferrándiz, Spanish National Research Council; Donald Boström, Investigative Journalist, Sweden; Michael Montgomery, Center for Investigative Reporting; Dr. Jorge Pérez Ávila, Former Director, Pedro Khori Institute, Havana; Meira Waiss, Professor Emerita of Antropology, Hebrew University, Jerusalem

 Gail Kligman, Assoc. Provost, Professor of Sociology, UCLA

 Nicholas B. Dirks, Professor of Anthropology, Chancellor, UCB

 Department of Anthropology, Center for Latin American Studies, Institute for the Study of Societal Issues, Humanities Research Institute, Center for Law and Society

The Anthropology department is hosting an international conference with special guests to honor the life and work of Professor Nancy Scheper-Hughes during May 1st and 2nd. Several of our own UCB Anthropology PhD alumni and other distinguished anthropologists, social scientists, doctors, epidemiologists, investigative journalists will also be speaking at the two day event.

String-Math Seminar: Geometry behind the shuffle conjecture

Seminar | May 1 | 2-3 p.m. | 402 LeConte Hall

 Erik Carlsson, UC Davis

 Department of Mathematics

The original "shuffle conjecture" of Haglund, Haiman, Loehr, Ulyanov, and Remmel predicted a striking combinatorial formula for the bigraded character of the diagonal coinvariant algebra in type A, in terms of some fascinating parking functions statistics. I will start by explaining this formula, as well as the ideas that went into my recent proof of this conjecture with Anton Mellit, namely the...   More >

Northern California Symplectic Geometry Seminar: Contact and symplectic foliations/Skeleta of Weinstein manifolds

Seminar | May 1 | 2:30-5 p.m. | Stanford University Mathematics Department, 384H/383N

 Alvaro del Pino/Laura Starkston, ICMAT, University of Madrid/Stanford University

 Department of Mathematics

In dimension 3, every foliation...   More >

Deepak Rajan - Distributed-memory branch-and-bound tree search for stochastic Mixed-Integer Programs (MIPs)

Seminar | May 1 | 3:30-5:30 p.m. | 3108 Etcheverry Hall

 Deepak Rajan, University of California Berkeley

 Industrial Engineering & Operations Research

In this talk, I present and compare two frameworks for distributed-memory branch-and-bound (B&B) tree search.

Physical Mechanisms of Cell Organization on Micron Length Scales

Seminar | May 1 | 4-5 p.m. | 106 Stanley Hall

 Michael Rosen, UT Southwestern Medical Center

 Department of Molecular and Cell Biology, Department of Chemistry

Student Algebraic Geometry Seminar: A conjecture in Diophantine geometry, via model theory

Seminar | May 1 | 4-5 p.m. | 891 Evans Hall

 Michael Wan, UC Berkeley

 Department of Mathematics

In 2011, Jonathan Pila published an unconditional proof of the André-Oort conjecture in Diophantine geometry, in the Annals. A simplified, vague form of the statement is as follows: if X is an irreducible affine complex variety, and X contains a Zariski dense set of “special points”, then X is a “special variety”.

Pila’s proof uses the theory of o-minimality from model theory, a...   More >

Seminar 271, Development: Mechanism Design meets Development: Selective Trials for Technology Diffusion

Seminar | May 1 | 4:10-5:30 p.m. | 648 Evans Hall

 Pascaline Dupas, Stanford

 Department of Economics

Dissertation Talk: Parallel Machine Learning Using Concurrency Control

Seminar | May 1 | 5-6 p.m. | 465H Soda Hall

 Xinghao Pan, EECS

 Electrical Engineering and Computer Sciences (EECS)

Many machine learning algorithms iteratively process datapoints and transform some global model parameters. It has become increasingly impractical to serially execute such iterative algorithms as processor speeds fail to catch up to the explosive growth in dataset sizes.

To address these problems, the machine learning community has turned to two parallelization strategies: bulk synchronous...   More >