Distinguished Lecture Series
2013-2014 Speakers
Ben Taskar
Boeing Associate Professor, Computer Science and Engineering
University of Washington
Probabilistic Models of Diversity: Determinantal Point Processes in Machine Learning
Tuesday, October 15, 2013
Abstract:
Determinantal point processes (DPPs) are elegant probabilistic models of repulsion that arise in quantum physics and random matrix theory. In contrast to traditional structured models like Markov random fields, which become intractable and hard to approximate in the presence of negative correlations, DPPs offer efficient and exact algorithms for sampling, marginalization, conditioning, and other inference tasks. We provide a gentle introduction to DPPs, focusing on the intuitions, algorithms, and extensions that are most relevant to the machine learning community, and show how DPPs can be applied to real-world applications like finding diverse sets of high-quality search results, building informative summaries by selecting diverse sentences from documents, modeling nonoverlapping human poses in images or video, and automatically building timelines of important news stories.
Bio:
Taskar’s research interests include machine learning, natural language processing and computer vision. He has been awarded the Sloan Research Fellowship, the NSF CAREER Award, and selected for the Young Investigator Program by the Office of Naval Research and the DARPA Computer Science Study Group. His work on structured prediction has received best paper awards at several conferences.
Nick McKeown
Kleiner Perkins, Mayfield, and Sequoia Professor,
Faculty Director of the Open Networking Research Center; Electrical Engineering and Computer Science
Stanford University
Software Defined Networks and Streamlining the Plumbing
Tuesday, November 12, 2013
Abstract:
One consequence of software defined networking (SDN) is that it clearly defines the role of the switches. Essentially, a switch only needs to do four things: (1) parse each packet header to find the fields it is interested in, then (2) match on those fields, yielding (3) rules to modify the packet, and finally (4) forward each packet to its next hop. Put another way, we can abstract packet forwarding as a set of "parse-match-action" processing steps, whether we are describing a simple Ethernet switch that plumbs servers together in a data center, or a router that aggregates enormous amounts of traffic together in a WAN. The "parse-match-action" forwarding model has two nice consequences that I'll describe in the talk. First, it tells us why programmable NPUs (network processors) were the wrong model for building packet switches. Instead of a sea of hard-to program Turing-complete processors, we just need a very fast, simple, RISC-like hardware pipeline that implements the "parse-match-action" model. I will describe such a design. Second, now we have an abstract model of forwarding, we can verify that the forwarding plane is doing what we originally intended. This allows us to bring in a vast swathe of formal verification and debugging techniques from other fields. I believe that in 20 years, we will look back and recognize this as the time network plumbing got dramatically simpler and less error prone.
Bio:
McKeown’s current research interests include software defined networks (SDN), how to enable more rapid improvements to the Internet infrastructure, and tools and platforms for networking research and teaching. He has co-founded a number of companies and the Open Networking Foundation (ONF) in 2011. McKeown is a member of the US National Academy of Engineering (NAE), a Fellow of the Royal Academy of Engineering (UK), Fellow of the IEEE and the ACM. He has been awarded the British Computer Society Lovelace Medal, the IEEE Kobayashi Computer and Communications Award and the ACM Sigcomm Lifetime Achievement Award.
Maja Mataric
Chan Soon-Shiong Chair, Computer Science, Neuroscience and Pediatrics;
Vice Dean for Research, Viterbi School of Engineering;
Founding Director, USC Center for Robotics and Embedded Systems; Director, USC Robotics Research Lab
University of Southern California
Robots Among Us? Socially Assistive Human-Robot Interaction
Tuesday, January 21, 2014
Abstract:
Socially assistive robotics (SAR) is a new field of intelligent robotics that focuses on developing machines capable of assisting users through social rather than physical interaction. The robot's physical embodiment is at the heart of SAR's effectiveness, as it leverages the inherently human tendency to engage with lifelike (but not necessarily humanlike or otherwise biomimetic) social behavior. People readily ascribe intention, personality, and emotion to robots; SAR leverages this engagement stemming from non-contact social interaction involving speech, gesture, movement demonstration and imitation, and encouragement, to develop robots capable of monitoring, motivating, and sustaining user activities and improving human learning, training, performance and health outcomes.
Human-robot interaction (HRI) for SAR is a growing multifaceted research area at the intersection of engineering, health sciences, neuroscience, social, and cognitive sciences. This talk will describe our research into embodiment, modeling and steering social dynamics, and long-term user adaptation for SAR. The research will be grounded in projects involving analysis of multi-modal activity data, modeling personality and engagement, formalizing social use of space and non-verbal communication, and personalizing the interaction with the user over a period of months, among others. The presented methods and algorithms will be validated on implemented SAR systems evaluated by human subject cohorts from a variety of user populations, including stroke patients, children with autism spectrum disorder, and elderly with Alzheimers and other forms of dementia.
Bio:
Mataric’s lab focuses on enabling robots to help people through social rather than physical assistance. Her research into socially assistive robotics is developing robot-aided therapies for autism, stroke rehabilitation, dementia, and obesity mitigation by developing algorithms for human-robot interaction that involve embodiment, social dynamics, and long-term adaptation. Among other honors, Mataric is a Fellow of the AAAS and IEEE, recipient of the Presidential Mentoring Award, the Okawa Foundation Award, NSF Career Award, MIT TR35 Innovation Award, and the IEEE Robotics and Automation Society Early Career Award.