William Wang (UCSB)

Event Date: 

Wednesday, October 11, 2017 - 3:30pm to 4:30pm

Event Date Details: 

Refreshments served at 3:15pm.

Event Location: 

  • Sobel Seminar Room
  • Department Seminar Series
Title: Deep Reinforcement Learning for Knowledge Graph Reasoning
 
Abstract: Deep reinforcement learning has achieved several important breakthroughs in the past few years, including Go and Atari games. However, for many other artificial intelligence problems, the action space is even larger than these games. In this talk, I will describe our recent work on learning to reason in large scale knowledge graphs (KGs). More specifically, I will describe a novel reinforcement learning framework for learning multi-hop relational paths: we use a policy-based agent with continuous states based on knowledge graph embeddings, which reasons in a KG vector space by sampling the most promising relation to extend its path. Our approach includes a reward function that takes accuracy, diversity, and efficiency into consideration. Experimentally, we show that our proposed method outperforms a path-ranking based algorithm and knowledge graph embedding methods on Freebase and Never-Ending Language Learning datasets, offering explainability and performance at the same time.
 
Bio: William Wang is an assistant professor in the Department of Computer Science at the University of California, Santa Barbara. He received his PhD from School of Computer Science, Carnegie Mellon University. He has broad interests in machine learning approaches to data science, including statistical relational learning, information extraction, computational social science, speech, and vision. He has published more than 40 papers at leading conferences and journals, and received best paper awards (or nominations) at ASRU 2013, CIKM 2013, and EMNLP 2015, a best reviewer award at NAACL 2015, an IBM faculty award in 2017, and the Richard King Mellon Presidential Fellowship in 2011. He is an alumnus of Columbia University, and a former research scientist intern of Yahoo! Labs, Microsoft Research Redmond, and University of Southern California.