This thesis presents a structural model of one aspect of science, the theory-driven discovery of empirical laws, in terms of the knowledge structures and reasoning processes that it involves; and describes a machine learning system designed to embody the major features of the model, called OZ, which is designed to investigate the transport properties of an unknown membrane separating two solutions. Inductive data-driven discovery is an important process in science, but takes place within very tightly constrained limits defined by theoretical reasoning. An explicit specification of the possible search space for a law is a <i>law framework</i>; this takes the form of a law with some undetermined parameters. Inductive law discovery is the search for the values of these free parameters. According to the model <i>informal qualitative models</i> (IQMs) describing the essential structural features of a physical system are used to guide the selection of appropriate variables for scientific law discovery, and the selection of an appropriate mathematical function for a law. Our analysis differs from previous work in machine discovery in stressing the importance of models of internal structure in scientific discovery. OZ comprises a domain independent control structure and a set of domain independent procedures, plus a set of domain dependent heuristics for the membrane properties domain. It constructs a set of candidate IQMs for the unknown membrane, and designs goal-directed experiments to determine which IQM is the right one, generating and testing qualitative predictions about the patterns to be expected in numerical data. When it has identified a single model as correct, it constructs law frameworks for possible laws describing the transport properties of the membrane, then designs different experiments to gather data to supply to an inductive law discovery function, which looks for a law of the type specified by each law framework.

No description available.

No description available.

No description available.

No description available.

No description available.

This research presents a novel technique termed Multi-Agent Malicious Behaviour Detection. The goal of Multi-Agent Malicious Behaviour Detection is to provide infrastructure to allow for the detection and observation of malicious multi-agent systems in computer network environments. This research explores combinations of machine learning techniques and fuses them with a multi-agent approach to malicious behaviour detection that effectively blends human expertise from network defenders with modern artificial intelligence. Success of the approach depends on the Multi-Agent Malicious Behaviour Detection system's capability to adapt to evolving malicious multi-agent system communications, even as the malicious software agents in network environments vary in their degree of autonomy and intelligence. This thesis research involves the design of this framework, its implementation into a working tool, and its evaluation using network data generated by an enterprise class network appliance to simulate both a standard educational network and an educational network containing malware traffic.

In this paper, I describe the application of genetic programming to evolve a controller for a robotic tank in a simulated environment.The purpose is to explore how genetic techniques can best be applied to produce controllers based on subsumption and behavior oriented languages such as REX. As part of my implementation, I developed TableRex, a modification of REX that can be expressed on a fixed-lengthgenome. Using a fixed subsumption architecture of TableRex modules, I evolved robots that beat some of the most competitive hand-coded adversaries.

In this paper we focus on the problem of estimating a boundeddensity using a finite combination of densities from a givenclass. We consider the Maximum Likelihood Procedure (MLE) and the greedy procedure described by Li and Barron. Approximation and estimation bounds are given for the above methods. We extend and improve upon the estimation results of Li and Barron, and in particular prove an $O(\frac{1}{\sqrt{n}})$ bound on the estimation error which does not depend on the number of densities in the estimated combination.

Previous studies have shown that dyslexic individuals who supplement windowed reading practice with intensive small-scale hand-eye coordination tasks exhibit marked improvement in their reading skills. Here we examine whether similar hand-eye coordination activities, in the form of artwork performed by children in kindergarten, first and second grades, could reduce the number of students at-risk for reading problems. Our results suggest that daily hand-eye coordination activities significantly reduce the number of students at-risk. We believe that the effectiveness of these activities derives from their ability to prepare the students perceptually for reading.