Seminar by Parag Singla
Markov Logic: Theory, Algorithms and Applications
Parag Singla
Department of Computer Science, University of Texas at Austin.
Date: Tuesday, May 10th, 2011
Time: 12:00 NOON
Venue: CS101.Abstract:
Most real world problems are characterized by relational structure i.e. entities and relationships between them. Further, they are inherently uncertain in nature. Logic gives the power to handle the first. Probability handles the later. Combining the two has been a long standing goal of AI as well as contemporary machine learning research. Markov logic achieves this by attaching real-valued weights to formulas in first-order logic. Formulas in Markov logic can be seen as defining the templates for constructing the ground Markov networks. Carrying out propositional inference techniques in such models leads to explosion in time and memory. To overcome these challenges, I will present lifted belief propagation (LBP), the first lifted approximate inference algorithm. LBP works by performing inference over an equivalent lifted network which can be potentially much smaller than the original ground network. I will then talk about two important applications (amongst many others) of Markov logic a) Entity Resolution b) Abductive Plan Recognition. I will conclude the talk with the directions for future work.
About the speaker:
Dr. Parag Singla is a post-doctoral researcher at U. Texas at Austin. He did his PhD in machine learning from Washington Univ. Seattle in 2009 and his BTech from IIT Bombay in 2002. His interests are Markov Logic and machine learning.