SUNMONTUEWEDTHUFRISAT
Events for April 12, 2018
-
EE Seminar: Towards Generalizable Imitation in Robotics
Thu, Apr 12, 2018 @ 01:30 PM - 02:30 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: Animesh Garg, Postdoctoral Researcher, Stanford University AI lab
Talk Title: Towards Generalizable Imitation in Robotics
Abstract: Robotics and AI are experiencing radical growth, fueled by innovations in data-driven learning paradigms coupled with novel device design, in applications such as healthcare, manufacturing and service robotics. And in our quest for general purpose autonomy, we need abstractions and algorithms for efficient generalization.
Data-driven methods such as reinforcement learning circumvent hand-tuned feature engineering, albeit lack guarantees and often incur a massive computational expense: training these models frequently takes weeks in addition to months of task-specific data-collection on physical systems. Further such ab initio methods often do not scale to complex sequential tasks. In contrast, biological agents can often learn faster not only through self-supervision but also through imitation. My research aims to bridge this gap and enable generalizable imitation for robot autonomy. We need to build systems that can capture semantic task structures that promote sample efficiency and can generalize to new task instances across visual, dynamical or semantic variations. And this involves designing algorithms that unify in reinforcement learning, control theoretic planning, semantic scene & video understanding, and design.
In this talk, I will discuss two aspects of Generalizable Imitation: Task Imitation, and Generalization in both Visual and Kinematic spaces. First, I will describe how we can move away from hand-designed finite state machines by unsupervised structure learning for complex multi-step sequential tasks. Then I will discuss techniques for robust policy learning to handle generalization across unseen dynamics. I will revisit structure learning for task-level understanding for generalization to visual semantics.
And lastly, I will present a program synthesis based method for generalization across task semantics with a single example with unseen task structure, topology or length. The algorithms and techniques introduced are applicable across domains in robotics; in this talk, I will exemplify these ideas through my work on medical and personal robotics.
Biography: Animesh is a Postdoctoral Researcher at Stanford University AI lab. Animesh is interested in problems at the intersection of optimization, machine learning, and design. He studies the interaction of data-driven Learning for autonomy and Design for automation for human skill-augmentation and decision support. Animesh received his Ph.D. from the University of California, Berkeley where he was a part of the Berkeley AI Research center and the Automation Science Lab. His research has been recognized with Best Applications Paper Award at IEEE CASE, Best Video at Hamlyn Symposium on Surgical Robotics, and Best Paper Nomination at IEEE ICRA 2015. And his work has also featured in press outlets such as New York Times, UC Health, UC CITRIS News, and BBC Click.
Host: Pierluigi Nuzzo, nuzzo@usc.edu
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 132
Audiences: Everyone Is Invited
Contact: Mayumi Thrasher
-
2018 Viterbi Keynote Lecture
Thu, Apr 12, 2018 @ 04:00 PM - 05:00 PM
Ming Hsieh Department of Electrical and Computer Engineering
Conferences, Lectures, & Seminars
Speaker: David Tse, Thomas Kailath and Guanghan Xu Professor, Stanford University
Talk Title: Maximum likelihood Genome Sequencing
Series: Viterbi Lecture
Abstract: Genome sequencing is one of the biggest breakthroughs in science in the past two decades. Modern sequencing methods use linking data at multiple scales to reconstruct pertinent information about the genome. Many such reconstruction problems can be formulated as maximum likelihood sequence decoding from noisy linking data. We discuss two in this talk: haplotype phasing, the problem of sequencing genomic variations on each of the maternal and paternal chromosomes, and genome scaffolding, the problem of finishing genome assembly using long-range 3D contact data. While maximum likelihood sequence decoding is NP-hard in both of these problems, spectral and linear programming relaxations yield efficient approximation algorithms that can provably achieve the information theoretic limits and perform well on real data. These results parallel the biggest success of information theory: efficiently achieving the fundamental limits of communication.
Biography: David Tse received the B.A.Sc. degree in systems design engineering from University of Waterloo in 1989, and the M.S. and Ph.D. degrees in electrical engineering from Massachusetts Institute of Technology in 1991 and 1994 respectively. From 1995 to 2014, he was on the faculty of the University of California at Berkeley. He received the Claude E. Shannon Award in 2017 and was elected member of the U.S. National Academy of Engineering in 2018. Previously, he received a NSF CAREER award in 1998, the Erlang Prize from the INFORMS Applied Probability Society in 2000 and the Frederick Emmons Terman Award from the American Society for Engineering Education in 2009. He received multiple best paper awards, and is the inventor of the proportional-fair scheduling algorithm used in all third and fourth-generation cellular systems.
Host: Sandeep Gupta, sandeep@usc.edu
More Info: https://minghsiehee.usc.edu/viterbi-lecture/
Webcast: https://bluejeans.com/401381224/More Information: 20180412 Tse Flyer.pdf
Location: Hughes Aircraft Electrical Engineering Center (EEB) - 132
WebCast Link: https://bluejeans.com/401381224/
Audiences: Everyone Is Invited
Contact: Mayumi Thrasher
Event Link: https://minghsiehee.usc.edu/viterbi-lecture/
