Thinking Ahead: Continual Computation Policies for Allocating Idle and Real-Time Resources to Solve Future Challenges

Eric Horvitz

Adaptive Systems and Interaction
Microsoft Research
Redmond, Washington 98052-6399

Access postscript or pdf file.

Abstract:

Research on continual computation centers on developing precomputation policies that can effectively harness available resources to solve future challenges. We focus on integrating a consideration of offline and real-time resources in continual computation. We review precomputation policies for flexible procedures and present strategies that account for the expected future real-time refinement of a result following precomputation. Finally, we address policies that consider the tradeoff between the efficiency of solving current and potential future challenges.

Keywords: Decision-theoretic control of computation, probability, decision theory, flexible computation, decisions under bounded resources, continual computation, bounded optimality.

In: Proceedings of IJCAI '99, Sixteenth Joint Conference on Artificial Intelligence, Stockholm, Sweden, July 1999. Morgan Kaufmann: San Francisco. pp. 1280-1286.

Author Email: horvitz@microsoft.com

Related Background

• Earlier paper on Models of Continual Computation (AAAI 97).
  
  
• Application of continual computation to decision-theoretic prefetching of web pages (CIKM 98).
  
  
• IEEE Intelligent Systems (March, April '98) article on research at Microsoft Research (.pdf file, read with Acrobat viewer; article copyright IEEE Intelligent Systems 1998, used with permission.).