Colloquium Details
Self-Adjusting Computation
Author: | Umut A. Acar Toyota Technological Institute, University of Chicago |
---|---|
Date: | May 06, 2008 |
Time: | 14:00-Time Change |
Location: | 220 Deschutes |
Host: | Yannis Smaragdakis |
Abstract
Self-adjusting computation refers to a model of computation where programs can respond to changes to their data automatically. The model has been motivated by applications from a diverse set of fields where data sets change dynamically over time. In this talk, I give an overview of the progress in developing the underpinnings of the model and in developing language and compiler support for writing self-adjusting programs. I consider a number of application domains (e.g., motion modeling, scientific computing, machine learning, computational biology) and show that self-adjusting programs can deliver good performance, both in theory and in practice.
Biography
Umut A. Acar is an Assistant Professor at Toyota Technological Institute at Chicago. He is broadly interested in language and algorithm design and implementation, particularly for dynamic systems that interact with changing data from various sources such as users and the physical environment. He has been working on the development of self-adjusting-computation model, where computations can respond to changes to their data (state) automatically. He received his Ph.D. from Carnegie Mellon University in 2005, his M.A. from University of Texas at Austin in 1999, and his B.S. from Bilkent University, Turkey, in 1997.