Location
CS Lobby
Event Description
Speaker: Alex Schwing - University of Toronto
Time: Tuesday Nov 24, 2.30-3.30pm
Location: New Computer Science building Rm 120
Reception is right after the talk

 

Title: Parallel Inference and Learning with Deep Structured Distributions
 
Abstract: Many problems in real-world applications involve predicting several
random variables which are statistically related. A structured model, like a
Markov random field, is a great mathematical tool to encode those dependencies.
 
Within the first part of this talk I will discuss the difficulties in finding
the most likely configuration described by a structured distribution. I will
present a model-parallel inference algorithm and illustrate its effectiveness in
jointly estimating the disparity of more than 12 million variables.
 
In the second part, I will show how to combine structured distributions with
deep learning to estimate complex representations which take into account the
dependencies between the random variables. To model those deep structured
distributions I will present a sample-parallel training algorithm and show its
applicability, among others, by using a 3D scene understanding task.
 
Bio: Alex Schwing studied electrical engineering and information technology at
Technical University of Munich (TUM) and completed his PhD at ETH Zurich, mainly
collaborating with Tamir Hazan (Technion), Raquel Urtasun (University of
Toronto), and Marc Pollefeys (ETH Zurich). He is currently a postdoctoral fellow
at University of Toronto. His research focuses on algorithms for inference and
learning of structured distributions, and his work is motivated among others by
applications arising from 3D scene understanding.
 
 
Hosted By
Minh Hoai Nguyen
Event Title
Fac Cand: Alex Schwing, University of Toronto