Conferences, Lectures, & Seminars
Events for December
-
Non-Parametric Latent Variable Models for Shape and Motion Analysis
Tue, Dec 02, 2008 @ 04:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Dr. Raquel Urtasun, MIT
Host: Prof. Fei Sha, Prof. Ram NevatiaAbstract:
Dimensionality reduction is a popular approach to dealing with high dimensional data sets. It is often the case that linear dimensionality reduction, such as principal component analysis (PCA), does not adequately capture the structure of the data. In this talk I will discuss Probabilistic Non-linear Latent Variable models in the context of 3D human body tracking, 3D shape recovery from single images, character animation and classification.First, I will describe how to use Gaussian Process Latent Variable models (GPLVMs) for learning human pose and motion priors for 3D human body tracking from monocular images. I will then show how to combine multiple local models to model the space of possible deformations of objects of arbitrary shapes, but made of the same material. This will allow us to perform monocular 3D shape recovery in the presence of complex deformations of poorly textured objects.In dimensionality reduction approaches, the data is typically embedded in a Euclidean latent space. However for some data sets, such as human motion, this is inappropriate. We present a range of approaches for embedding data into non-Euclidean latent spaces that incorporate prior knowledge. This allows us to learn models suitable for motion generation with good generalization properties. Finally, I.ll show how to extend these models to be discriminative, resulting in accurate classification even when dealing with high dimensional input spaces and very small training sets.Biography:
Raquel Urtasun is currently a Research Scientist at the International Computer Science Institute at Berkeley, working with Prof. Trevor Darrell and will be a Visiting Scholar at UC Berkeley EECS. Raquel’s main research areas are computer vision, machine learning and computer graphics. During 2006-2008, she was a postdoctoral associate at MIT-CSAIL. She earned her PhD at EPFL (Switzerland) in 2006 under the supervision of Prof. Pascal Fua on Motion Models for Robust 3D Human Body Tracking.
Location: Charles Lee Powell Hall (PHE) - 223
Audiences: Everyone Is Invited
Contact: CS Colloquia
-
Autonomy: From Outer to Inner Space
Thu, Dec 11, 2008 @ 04:00 PM
Thomas Lord Department of Computer Science
Conferences, Lectures, & Seminars
Speaker: Dr. Kanna Rajan, Principal Researcher for Autonomy, Monterey Bay Aquarium Research Institute (MBARI)
Host: Prof. Gaurav SukhatmeAbstract:
Ocean Sciences the world over is at a cusp, with a move from the Expeditionary to the Observatory mode of doing science. Recent policy decisions in the United States, are pushing the technology for persistent observation and sampling which hitherto had been either economically unrealistic or unrealizable due to technical constraints. With the advent of ocean observatories, a number of key technologies have however proven to be promising for sustained ocean presence. Mobile robots routinely map the benthic environment and sample the water-column up to depths of 6000 meters while tele-operated vehicles navigate remote depths performing scientific experiments in-situ relating to biogeochemical processes in the ocean. Such platforms however have inherent limitations with how they are commanded and operated; pre-defined sequences of commands are currently used to determine what actions the robot will perform and when irrespective of the contextual environment in which it operates. As a consequence not only can the robot not recover from unforeseen failure conditions, but they are unable to significantly leverage their substantial onboard assets to do opportunistic science.To mitigate such shortcomings, we are developing deliberative techniques to dynamically command Autonomous Underwater Vehicles (AUV). Our effort is aimed to use a blend of generative and deliberative Artificial Intelligence Planning and Execution techniques to shed goals, introspectively analyze onboard resources and recover from failures. In addition we are working on clustering techniques to adaptively trigger science instruments that will contextually sample the seas driven by scientific intent. The end goal is towards unstructured exploration of the subsea environments that are a rich trove of problems for autonomous systems. Our work is a continuum of efforts from research at NASA to command deep space probes and Mars rovers, the lessons of which we have factored into the oceanic domain. In this talk I will articulate the challenges of working in the hostile underwater domain, lay out the differences and motivate our architecture for goal-driven autonomy on AUVs.Biography:
Dr. Kanna is the Principal Researcher in Autonomy at the Monterey Bay Aquarium Research Institute (http://www.mbari.org) a privately funded non-profit Oceanographic institute which he joined in October 2005. Prior to that he was a Senior Research Scientist and a member of the management team of the the 95 member Autonomous Systems and Robotics Area at NASA Ames Research Center Moffett Field, California. At Ames, balanced programmatic and technical responsibilities. He was the Principal Investigator of the MAPGEN Mixed-Initiative Planning effort to command and control the Spirit and Opportunity rovers on the surface of the Red Planet. MAPGEN continues to be used to this day, twice daily in the mission-critical uplink process at the Jet Propulsion Laboratory in Pasadena. Kanna was one of the six principals of the Remote Agent Experiment (RAX) team, which designed, built, tested and flew the first closed-loop AI based control system on a spacecraft. The RA was the co-winner of NASA's 1999 Software of the Year, the agency's highest technical award (http://ic.arc.nasa.gov/projects/remote-agent/). His interests are in automated Planning/Scheduling, modeling and representation for real world planners and agent architectures for Distributed Control applications. Prior to joining NASA Ames, he was in the doctoral program at the Courant Institute of Math Sciences at NYU. Prior to that he was at the Knowledge Systems group at American Airlines, helping build a Maintenance Routing scheduler (MOCA), which continues to be used by the airline 365 days of the year. MAPGEN has been awarded NASA's 2004 Turning Goals into Reality award under the Administrators Award category, a NASA Space Act Award, a NASA Group Achievement Award and a NASA Ames Honor Award. Kanna is the recipient of the 2002 NASA Public Service Medal and the First NASA Ames Information Directorate Infusion Award also in 2002. In Oct 2004, JPL awarded him the NASA Exceptional Service Medal for his role on the Mars Exploration Rovers misson. He was the Co-chair of the 2005 Intnl. Conference on Automated Planning and Scheduling (ICAPS), Monterey California (http://icaps05.icaps-conference.org/) and till recently the chair of the Executive Board of the International Workshop on Planning and Scheduling for Space. He continues to serve on review panels for NASA, the Italian Space Agency and European Space Agency.Location: Seaver Science Library (SSL) - 150
Audiences: Everyone Is Invited
Contact: CS Colloquia
