Home > Xiangliang Zhang - Walk-to-Win: Training Smart Agents for Attributed Network Node Classification

Xiangliang Zhang - Walk-to-Win: Training Smart Agents for Attributed Network Node Classification

Start:

2/6/2020 at 3:30PM

End:

2/6/2020 at 4:45PM

Location:

126 DeBartolo

Host:

College of Engineering close button
headerbottom

Nitesh Chawla

Nitesh Chawla

VIEW FULL PROFILE Email: nchawla@nd.edu
Phone: 574-631-1090
Website: https://niteshchawla.nd.edu
Office: 384 Nieuwland Science Hall

Affiliations

College of Engineering Frank M. Freimann Professor
Chawla's research interests are broadly in the areas of Big Data: data science, machine learning, network science and their applications social networks, healthcare informatics/analytics, and climate/environmental sciences.
Click for more information about Nitesh
574-631-1090
Add to calendar:
iCal vCal

Classification of nodes in an attributed network faces challenges of the integration of network structure and node text information, the heterogeneity of vertex attributes, and the lack of sufficient labeled nodes.  This talk will introduce our recent solutions to this problem based on reinforcement learning for intelligently aggregating a vertex’s neighborhood information. We train agents (decision-makers) to walk on a graph to win a game of node classification.  Agents get a reward if classifying correctly (by walking to the correct neighbors and collecting the correct information) and a penalty otherwise.  Agents keep learning to receive more rewards and make more correct decisions about where to walk. In addition, the walking path made by agents can be used to interpret the decision making process and infer class label dependency. We will show that this idea of “walk-to-win” is efficient on large-scale networks, and outperforms all state-of-the-art methods designed for the same problem.

Seminar Speaker:

Xiangliang Zhang

Xiangliang Zhang

King Abdullah University of Science and Technology

Dr. Xiangliang Zhang is an Associate Professor of Computer Science and directs the MINE (http://mine.kaust.edu.sa) group at KAUST, Saudi Arabia. She earned her Ph.D. degree in computer science from INRIA-Universite Paris-Sud, France, in July 2010. She received her M.S. and B.S. degrees from Xi’an Jiaotong University, China, in 2006 and 2003, respectively. Dr. Zhang's research mainly focuses on learning from complex and large-scale streaming data and graph data. Dr. Zhang has published over 130 research papers in referred international journals and conference proceedings, including TKDE, SIGKDD, AAAI, IJCAI, ICDM, VLDB J, ICDE etc.   She regularly serves on the Program Committee for premier conferences like SIGKDD (Senior PC), AAAI (Senior PC), IJCAI (Senior PC), ICDM, NIPS, ICML etc.  Dr. Zhang was invited to deliver an Early Career Spotlight talk at IJCAI-ECAI 2018.