# Colloquium

## 3D Scanning in Egypt and Image-based Object Editing

## Computers and the Sociology of Mathematical Proof

After reviewing briefly the history of these developments, the talk will explain why they are of interest to the sociology of scientific knowledge: they suggest the emergence of a set of "cultures of proving" different from those of ordinary mathematics. Clashes between cultures of automated proving and those of ordinary mathematics, and the first litigation centering on the meaning of mathematical "proof," will be examined. The talk will also discuss areas of tension within automated proving and outline a case in which the priorities of national security can be seen in the very construction of an automated theorem prover.

## Adiabatic quantum algorithms

## Statistical Analysis of Anatomical Shape and Function

## Computer Science and Game Theory: Mutual Influences and Synergies

- modeling the empirical computational hardness of selecting the winners of a combinatorial auction, analyzing these models, and applying them to construct algorithm portfolios and to generate hard problem instances
- using incentives to diffuse temporally-focused usage of network resources at the lowest possible cost
- local-effect games, a novel class of multi-player general-sum games for which representation is compact, pure-strategy Nash equilibria often exist, and equilibria can often be computed efficiently

## Genes,Tumors, and Bayes nets: Improving the specificity of biological signal detection in microarray analysis.

## Multiagent Learning in the Presence of Limited Agents

In this talk I focus on the contributions of the WoLF principle and the GraWoLF algorithm. I show that the WoLF variable learning rate causes learning to converge to optimal responses in settings of simultaneous learning. I demonstrate this converging effect both theoretically in a subclass of single-state games and empirically in a variety of multiple-state domains. I then describe GraWoLF, a combination of policy gradient techniques and the WoLF principle. I show compelling results of applying this algorithm to a card game with an intractably large state space as well as an adversarial robot task. These results demonstrate that WoLF-based algorithms can effectively learn in the presence of other learning agents, and do so even in complex tasks with limited agents.

## Efficiency in Online and Noise-Tolerant Learning

## Multiagent Planning with Factored MDPs

This talk presents a framework for approximate planning that can exploit structure in a factored MDP to solve problems with many trillions of states and actions. The talk will focus on three key elements:

- Factored Linear Programs -- A novel LP decomposition technique, using ideas from inference in Bayesian networks, which can yield exponential reductions in planning time.
- Distributed Coordination -- A distributed multiagent decision making algorithm, where the coordination structure arises naturally from the system dynamics.
- Generalization in Relational MDPs -- A method for learning general solutions from solved tasks, that allows us to act in new scenarios without replanning.

We demonstrate our approach on the task of multiagent coordination in a real strategic computer war game.

## Neptune: Programming and Runtime Support for Cluster-based Network Services

Neptune is a middleware system that allows services to be aggregated and deployed quickly on a cluster of workstations and SMPs. It shields application programmers from the complexities of replication, service discovery, failure detection and recovery, and resource management. The techniques investigated are architecture and programming support for thread/process based concurrency, quality-aware service differentiation, parallel data aggregation, and replica consistency management with staleness control.

This talk will also discuss the use of Neptune in Internet search at Ask Jeeves and Teoma.

Tao Yang is an Associate Professor of Computer Science at University of California at Santa Barbara. He has also been the Chief Scientist for Teoma and Ask Jeeves since 2000 for directing research and development of Internet search engines. His research interests are parallel and distributed systems, high performance scientific computing, and Internet search.