CSL Master Calendar

Back to Listing

DCL Seminar: Aram Galstyan - Generative Models for Multimodal Network Data

Event Type
Decision and Control Laboratory, Coordinated Science Laboratory
CSL Auditorium, Room B02
Nov 15, 2017   3:00 pm  
Aram Galstyan, Ph.D., Information Sciences Institute, University of Southern California
Linda Stimson
Originating Calendar
CSL Decision and Control Group

Decision and Control Lecture Series

Coordinated Science Laboratory


 “Generative Models for Multimodal Network Data”


Aram Galstyan, Ph.D.

Information Sciences Institute

University of Southern California


Wednesday, November 15, 2017

3:00 p.m. to 4:00 p.m.

CSL Auditorium (B02)



Studies of social systems have traditionally focused on analyzing networks induced by social interactions, while discarding rich contextual information on nodes and their properties. At the same time, empirical evidence points to strong correlations between node attributes and their interactions.  Here we suggest an efficient generative framework for analyzing attribute-rich social data. Within this approach, each node is assigned an unobserved (latent) position in some space, so that both the nodes’ attributes and their interactions depend on their coordinates in this space. This “shared” latent space allows to capture observed correlations between the attributes and network structure. We perform extensive experiments where the goal is predict missing links in a network using attributes, or predict user attributes based on network information, and observe that the proposed method outperforms other baselines in both prediction tasks. We also consider temporal extension of  the generative model and demonstrate its usefulness in detecting both local and global changes in evolving networks. 


Dr. Aram Galstyan is Research Director for Machine Intelligence and Data Science at the Information Sciences Institute, University of Southern California, and Research Associate Professor in the USC Computer Science Department. Dr. Galstyan’s current research focuses on various problems at the intersection of machine learning, information theory, and statistical physics. His research includes both theoretical effort and more application-oriented work geared toward describing various real-world phenomena. His research has been supported by various U.S. funding agencies, including NSF, NIH, DARPA, IARPA, and ARO.

link for robots only