Explorations of the Practical Issues of Learning Prediction-Control Tasks Using Temporal Difference Learning Methods

There has been recent interest in using temporal difference learning methods to attack problems of prediction and control. While these algorithms have been brought to bear on many problems, they remain poorly understood. It is the purpose of this thesis to further explore these algorithms, presenting a framework for viewing them and raising a number of practical issues and exploring those issues in the context of several case studies. This includes applying the TD(lambda) algorithm to: 1) learning to play tic-tac-toe from the outcome of self-play and of play against a perfectly-playing opponent and 2) learning simple one-dimensional segmentation tasks.

The Informational Complexity of Learning from Examples

This thesis attempts to quantify the amount of information needed to learn certain tasks. The tasks chosen vary from learning functions in a Sobolev space using radial basis function networks to learning grammars in the principles and parameters framework of modern linguistic theory. These problems are analyzed from the perspective of computational learning theory and certain unifying perspectives emerge.

Task-Level Robot Learning

We are investigating how to program robots so that they learn from experience. Our goal is to develop principled methods of learning that can improve a robot's performance of a wide range of dynamic tasks. We have developed task-level learning that successfully improves a robot's performance of two complex tasks, ball-throwing and juggling. With task- level learning, a robot practices a task, monitors its own performance, and uses that experience to adjust its task-level commands. This learning method serves to complement other approaches, such as model calibration, for improving robot performance.

Learning from Ambiguity

There are many learning problems for which the examples given by the teacher are ambiguously labeled. In this thesis, we will examine one framework of learning from ambiguous examples known as Multiple-Instance learning. Each example is a bag, consisting of any number of instances. A bag is labeled negative if all instances in it are negative. A bag is labeled positive if at least one instance in it is positive. Because the instances themselves are not labeled, each positive bag is an ambiguous example. We would like to learn a concept which will correctly classify unseen bags. We have developed a measure called Diverse Density and algorithms for learning from multiple-instance examples. We have applied these techniques to problems in drug design, stock prediction, and image database retrieval. These serve as examples of how to translate the ambiguity in the application domain into bags, as well as successful examples of applying Diverse Density techniques.

Representation and Detection of Shapes in Images

We present a set of techniques that can be used to represent and detect shapes in images. Our methods revolve around a particular shape representation based on the description of objects using triangulated polygons. This representation is similar to the medial axis transform and has important properties from a computational perspective. The first problem we consider is the detection of non-rigid objects in images using deformable models. We present an efficient algorithm to solve this problem in a wide range of situations, and show examples in both natural and medical images. We also consider the problem of learning an accurate non-rigid shape model for a class of objects from examples. We show how to learn good models while constraining them to the form required by the detection algorithm. Finally, we consider the problem of low-level image segmentation and grouping. We describe a stochastic grammar that generates arbitrary triangulated polygons while capturing Gestalt principles of shape regularity. This grammar is used as a prior model over random shapes in a low level algorithm that detects objects in images.

Learning from Incomplete Data

Real-world learning tasks often involve high-dimensional data sets with complex patterns of missing features. In this paper we review the problem of learning from incomplete data from two statistical perspectives---the likelihood-based and the Bayesian. The goal is two-fold: to place current neural network approaches to missing data within a statistical framework, and to describe a set of algorithms, derived from the likelihood-based framework, that handle clustering, classification, and function approximation from incomplete data in a principled and efficient manner. These algorithms are based on mixture modeling and make two distinct appeals to the Expectation-Maximization (EM) principle (Dempster, Laird, and Rubin 1977)---both for the estimation of mixture components and for coping with the missing data.