Efficient Computational Methods with Provable Guarantees for Data-Driven Problems

3108 Etcheverry Hall
  • Somayeh Sojoudi, Berkeley IEOR and EECS
The area of data science lacks efficient computational methods with provable guarantees that can cope with the large-scale nature and the high nonlinearity of many real-world systems. Practitioners often design heuristic algorithms tailored to specific applications, but the theoretical underpinnings of these methods remain a mystery and this limits their usage in safety-critical systems. In this...

The Limits of Proof

Banatao Auditorium Sutardja Dai Hall
  • Paul Beame, University of Washington
In the early part of the 20th century, Gödel, Turing, and Tarski showed that no consistent system of reasoning can contain proofs of important properties of the natural numbers or of computations. In these cases, the difficulty stems from the need to reason about infinities of numbers or time that don't show up in our everyday world. In contrast, proofs of properties in a bounded size world...

Dissertation Talk: Learning to Navigate in Visual Environments

380 Soda Hall
In this talk, we examine learning and systems challenges related to visual navigation tasks. First, we approach visual navigation tasks within the setting of reinforcement learning under partial observation. Second, we consider issues related to the large scale training of convolutional nets.

Design Field Notes: Ian Leighton

220 Jacobs Hall
Ian Leighton is a digital product designer from the Bay Area, now based in Lisbon. He graduated from UC Berkeley in Mechanical Engineering (2011). He founded a startup, Embark which created a dozen public transit apps for iPhone, and was acquired by Apple in 2013. At Apple, Ian worked on a large part of the design for Apple Maps Transit. His talk will explore "Design in Transit."

Creating the future of nuclear energy: Berkeley Distinguished Lectures in Data Science with Rachel Slaybaugh

190 Doe Library
  • Rachel Slaybaugh, Assistant Professor, Nuclear Engineering, UC Berkeley
The nuclear energy industry is at a crossroads: existing nuclear reactors are struggling to operate economically in some tough markets, and construction of new designs in the U.S. is slow and over budget. At the same time, interest in and development of the next generation of nuclear reactors is growing at an unprecedented rate, and some other nations are building new reactors efficiently. Can...

EECS Colloquium: Deep Learning to Solve Challenging Problems

306 (HP Auditorium) Soda Hall
  • Jeff Dean, Google, Inc.
For the past seven years, Google Research teams have conducted research on difficult problems in artificial intelligence, on building large-scale computer systems for machine learning research, and, in collaboration with many teams at Google, on applying our research and systems to many Google products. As part of our work in this space, we have built and open-sourced the TensorFlow system...

Rethinking America’s 20th-Century Highway InstitutionsModeling Dynamic Transportation Networks Using Differential Complementarity Systems

290 Hearst Memorial Mining Building
  • Xuegang (Jeff) Ban, University of Washington
University of Washington's Xuegang (Jeff) Ban will present Modeling Dynamic Transportation Networks Using Differential Complementarity Systems on November 30, 2018 at 4 p.m. in 290 Hearst Memorial Mining Building. Join us for cookies and beverages at 3:30 p.m.

Maps of a rising water table: The hidden component of sea level rise: Berkeley Distinguished Lectures in Data Science

190 Doe Library
  • Kristina Hill, Associate Professor, Landscape Architecture & Environmental Planning, UC Berkeley
Map-based data viewers have been available for several years that reveal where coastal flooding is likely to occur as oceans warm and ice sheets melt. Recently, geologists have begun to study the influence of sea level rise on groundwater, and have concluded that in some coastal areas, as much or more land could flood as a result of rising groundwater than will flood directly from saltwater. Yet...