Browsing by Author "Parker, Austin"
Now showing 1 - 2 of 2
Results Per Page
Sort Options
Item Spatial Probabilistic Temporal Databases(2008-09-10) Parker, Austin; Subrahmanian, VS; Computer Science; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Research in spatio-temporal probabilistic reasoning examines algorithms for handling data such as cell phone triangulation, GPS systems, movement prediction software, and other inexact but useful data sources. In this thesis I describe a probabilistic model theory for such data. The Spatial PrObabilistic Temporal database framework (or SPOT database framework) provides methods for interpreting, checking consistency, automatically revising, and querying such databases. This thesis examines two different semantics within the SPOT framework and presents polynomial-time consistency checking algorithms for both. It introduces several revision techniques for repairing inconsistent databases and compares them to the AGM Axioms for belief state revision; finding an algorithm that, by only changing the probability bounds in the SPOT atoms, can repair a SPOT database in polynomial time while still satisfying the AGM axioms. Also included is an investigation into optimistic and cautious versions of a selection query that returns all objects in a given region with at least (or at most) a certain probability. For these queries, I introduce an indexing structure akin to the R-tree called a SPOT tree, and show experiments where indexing speeds up selection with both artificial and real-world data. I also introduce query preprocessing techniques that bound the sets of solutions with both circumscribing and inscribing regions, and discover these to also provide query time improvements in practice. By covering semantics, consistency checking, database revision, indexing, and query preprocessing techniques for SPOT database, this thesis provides a significant step towards a SPOT database framework that may be applied to the sorts of real-world problems in the impressive amount of semi-accurate spatio-temporal data available today.Item Theoretical and Experimental Analysis of an Evolutionary Social-Learning Game(2012-01-13) Carr, Ryan; Raboin, Eric; Parker, Austin; Nau, DanaAn important way to learn new actions and behaviors is by observing others, and several evolutionary games have been developed to investigate what learning strategies work best and how they might have evolved. In this paper we present an extensive set of mathematical and simulation results for Cultaptation, which is one of the best-known such games. We derive a formula for measuring a strategy's expected reproductive success, provide algorithms to compute near-best-response strategies and near-Nash equilibria, and provide techniques for efficient implementation of those algorithms. Our experimental studies provide strong evidence for the following hypotheses: 1. The best strategies for Cultaptation and similar games are likely to be conditional ones in which the choice of action at each round is conditioned on the agent's accumulated experience. Such strategies (or close approximations of them) can be computed by doing a lookahead search that predicts how each possible choice of action at the current round is likely to affect future performance. 2. Such strategies are likely to exploit most of the time, but will have ways of quickly detecting structural shocks, so that they can switch quickly to innovation in order to learn how to respond to such shocks. This conflicts with the conventional wisdom that successful social-learning strategies are characterized by a high frequency of innovation; and agrees with recent experiments by others on human subjects that also challenge the conventional wisdom.