Title
Probabilistic Reasoning on Lie Groups with Application to Nonparametric Object
and Parts Modeling.

Abstract
Articulated motion analysis often utilizes strong prior knowledge such as a
known or trained parts model for humans. Yet, the world contains a variety of
articulating objects--mammals, insects, mechanized structures--where the number
and configuration of parts for a particular object is unknown in advance. Here,
we relax such strong assumptions via an unsupervised, Bayesian nonparametric
parts model that infers an unknown number of parts with motions coupled by a
body dynamic and parameterized by SE(D), the Lie group of rigid transformations.
We derive an inference procedure that utilizes short observation sequences
(image, depth, point cloud or mesh) of an object in motion without need for
markers or learned body models. Novel and efficient Gibbs decompositions for
inference over distributions on SE(D) demonstrate robust part decompositions of
moving objects under both 3D and 2D observation models. The inferred
representation permits new analysis, such as object segmentation by relative
part motion, and transfers to new observations of the same object type. Although
this talk focuses on SO(D) and SE(D), we introduce probabilistic reasoning over
general matrix Lie groups.

Bio
David's research enables scientists to collect, analyze and react to
observations at scale. By working towards automated interventions, he aims to
make new experimental designs possible from what once were observational
studies.

Broadly, his research interests are in interpretable machine learning and
computer vision, with special focus on Bayesian nonparametrics applied to
time-series, distributions on manifolds and using uncertainty to guide decision
making, analysis and experimental design.

David is a PhD student with John Fisher at CSAIL, MIT. He has won awards or
worked with Google, Microsoft, NASA and the Industrial Designers Society of
America. He has received the NSF Graduate Research Fellowship and his work has
led to grants from the NSF and NIH. He has been interviewed by national outlets
including NPR and Wired and once upon a time founded Essistive
through which he licensed assistive technologies.