Electrical Engineering and Computer Science

AI Seminar

Implications of Algorithmic and High-Frequency Trading and From Perception to Cognition: Towards Knowledge Driven Image Understanding

Michael Wellman and Jia Deng

Computer Science & Engineering, University of Michigan
Tuesday, October 28, 2014
4:00pm - 5:30pm
3725 BBB

Add to Google Calendar
This week there will be two presentations at our seminar.

About the Event

Separating the beneficial from potentially harmful effects of high-frequency trading requires distinguishing among roles and strategies of algorithmic traders. In previous work, we focused on latency arbitrage in fragmented markets, and found a negative influence of this high-frequency trading activity on market efficiency. In current work, we are studying algorithmic strategies for market making (MM). The MM simultaneously maintains offers to buy and sell in a two-sided market, providing liquidity in an effort to profit from price spreads. Liquidity is generally considered to improve market performance, but does the benefit accrue to background investors?
How do you recognize a Cardigan Welsh Corgi? Why does a cliff look dangerous? Why are you surprised to see a penguin in a desert? When you understand images, you make extensive use of external knowledge, that is, commonsense and/or domain expertise that is not directly visible. In this talk I'll present some recent work that explores the intersection of perception and cognition, in particular, how to leverage external knowledge to help visual learning. One project is on extracting knowledge from the crowd to recognize objects in specialized domains. Another is on using a knowledge graph to improve large-scale object classification.


Michael Wellman is a member of the University of Michigan AI Laboratory.
Jia Deng is an Assistant Professor of Computer Science and Engineering at the University of Michigan. He received his Ph.D. from Princeton University and his B.Eng. from Tsinghua University, both in computer science. His research in computer vision focuses on image and video understanding through big visual data, human computation, and large-scale machine learning. His work has been featured in popular press such as the New York Times. He is a recipient of the Yahoo ACE Award, the ICCV Marr Prize, and the ECCV Best Paper Award.

Additional Information

Sponsor(s): Toyota

Open to: Public