Trajectories through Knowledge Space: A Dynamic Framework for Machine Comprehension(286 The Springer International Series in Engineering and Computer Science)

"Trajectories through Knowledge Space: A Dynamic Framework for Machine Comprehension" provides an overview of many of the main ideas of connectionism (neural networks) and probabilistic natural language processing. Several areas of common overlap between these fields are described in which each community can benefit from the ideas and techniques of the other. The author's perspective on comprehension pulls together the most significant research of the last ten years and illustrates how we can move forward onto the next level of intelligent text-processing systems. A central focus of the book is the development of a framework for comprehension connecting research themes from cognitive psychology, cognitive science, corpus linguistics and artificial intelligence. The book proposes a new architecture for semantic memory, providing a framework for addressing the problem of how to represent background knowledge in a machine. This architectural framework supports a computational model of comprehension. "Trajectories through Knowledge Space: A Dynamic Framework for Machine Comprehension" should be a useful reference for researchers and professionals, and may be used as an advanced text for courses on the topic.

Table of Contents:
1 Introduction.- 1.1 Motivation.- 1.2 A View of Text Comprehension.- 1.3 Overview of the LeMICON System.- 1.4 Implementation.- 1.5 Points of Interest.- 1.6 The Impact of this Work on Four Related Fields.- 1.7 Development of the Two-Tier Model.- 1.8 A Guide to the Reader.- 2 An Overview of Connectionist and Probabilistic Approaches to Language Processing.- 2.1 A Computational Linguistics Perspective.- 2.2 A Connectionist Perspective.- 2.3 A Comparison of Connectionist and Probabilistic NLP Methods.- 2.4 Bridging the Gap — Integrating Connectionist and Computational Linguistic Approaches.- 3 Memory Architecture.- 3.1 The Relational Tier.- 3.2 The Associational or ASF Tier.- 3.3 Connecting the Two Tiers.- 3.4 Working Memory.- 3.5 How New Knowledge Is Integrated.- 4 The Basic Computation.- 4.1 A Functional Description of the Algorithm.- 4.2 Representing the Input.- 4.3 Computational Details and Program Output at each Step.- 4.4 General Discussion of the Algorithm.- 4.5 How LeMICON Handles Binding.- 4.6 The Links to Psychology and Neurophysiology Revisited.- 4.7 Some Comparisons to Other Text Understanding Systems.- 5 Analysis of the Interpretation at the Relational and ASF Level.- 5.1 Introduction.- 5.2 Analyzing the Interpretation at the Relational Level.- 5.3 Analyzing the Interpretation at the ASF Level.- 5.4 Analyzing Time-Dependent Interactions at the ASF Level.- 5.5 Comparing Interpretations — A Quantitative Analysis.- 5.6 An Ablation Study.- 6 Reasoning from the Relational Level of the Representation.- 6.1 Introduction.- 6.2 Identifying the Conceptual Roots.- 6.3 Explaining the Connections Between Events.- 6.4 Determining Important Concepts in the Interpretation Graph.- 6.5 Conceptual Roots and their Role in Summarization.- 7 Experiments in Acquiring Knowledgefrom On-line Corpora.- 7.1 Introduction.- 7.2 The Automatic Acquisition of Knowledge from On-line Sources.- 7.3 The Automatic Construction of the Relational Tier.- 7.4 The Automatic Construction of the Associational Tier.- 7.5 How Semantic Memory Evolves in Response to New Input.- 7.6 Changing the Link Weights.- 7.7 Implementation Details.- 8 An Analysis of the Acquired Knowledge.- 8.1 An Alternative View of the Underlying Knowledge Representation.- 8.2 Discussion of Soundness of Approach.- 8.3 An Evaluation of LeMICON’s Representation.- 8.4 Previous Text Systems Revisited.- 8.5 The Knowledge Acquisition Continuum.- 9 Conclusions.- 9.1 Some Consequences of the Two-Tier Model of Memory.- 9.2 Associational Representations.- 9.3 The Universality of ASFs.- 9.4 Scalability.- 9.5 Automatic Acquisition of Knowledge.- 9.6 Building Large-Scale Knowledge Bases.- 9.7 A Link to Corpus Linguistics.- 9.8 The Interplay between Computation and Representation.- 9.9 Limitations.- 10 Future Directions.- 10.1 Expanding The Knowledge Base.- 10.2 Finding Deeper Semantic Relationships via Corpus Analysis.- 10.3 Handling Contradictory Input.- 10.4 Learning New Relationships.- 10.5 A Basic Level Semantic Encoding.- 10.6 Child Versus Adult Comprehension.- References.- Author Index.