Drawing on ideas from cognitive linguistics, connectionism, and perception, The Human Semantic Potential describes a connectionist model that learns perceptually grounded semantics for natural language in spatial terms. Languages differ in the ways in which they structure space, and Regier's aim is to have the model perform its learning task for terms from any natural language. The system has so far succeeded in learning spatial terms from English, German, Russian, Japanese, and Mixtec. The model views simple movies of two-dimensional objects moving relative to one another and learns to classify them linguistically in accordance with the spatial system of some natural language. The overall goal is to determine which sorts of spatial configurations and events are learnable as the semantics for spatial terms and which are not. Ultimately, the model and its theoretical underpinnings are a step in the direction of articulating biologically based constraints on the nature of human semantic systems. Along the way Regier takes up such substantial issues as the attraction and the liabilities of PDP and structured connectionist modeling, the problem of learning without direct negative evidence, and the area of linguistic universals, which is addressed in the model itself. Trained on spatial terms from different languages, the model permits observations about the possible bases of linguistic universals and interlanguage variation. Neural Network Modeling and Connectionism series

Chapter 1 Introduction 1.1 Matter and Method 1.2 Space and Semantic Potential 1.3 Negative Evidence and Language Learning 1.4 The Modeling Challenge 1.5 Constrained Connectionism 1.6 A Brief Exchange 1.7 An Overview of the Book Chapter 2 The Linguistic Categorization of Space 2.1 Perception and the Human Semantic Potential 2.2 The Primacy of Space 2.3 Crosslinguistic Variation 2.4 Cognitive Linguistics 2.5 Summary Chapter 3 Connectionism and Cognitive Models 3.1 Overview 3.2 Parallel Distributed Processing 3.3 Structured Connectionism 3.4 Constrained Connectionism 3.5 Learning Sequences Using Back-Propagation 3.6 Another Brief Exchange Chapter 4 Learning without Explicit Negative Evidence 4.1 The Problem 4.2 A Solution: Mutual Exclusivity 4.3 Difficulties with Mutual Exclusivity 4.4 Salvaging Mutual Exclusivity 4.5 Implementation 4.6 Results Chapter 5 Structures 5.1 Structures and Constrained Connectionism 5.2 Orientation Combination 5.3 Map Comparison 5.4 Motion: Source, Path, and Destination Chapter 6 A Model of Spatial Semantics 6.1 Overview 6.2 The Problem 6.3 The Model 6.3.1 Movies as Input, Categorization as Output 6.3.2 A Constrained Connectionist Architecture 6.4 A Run through the Model 6.5 Results 6.6 Prototype Effects 6.7 Discussion Chapter 7 Extensions 7.1 Polysemy 7.2 Deixis 7.3 Prelinguistic Conceptual Development 7.4 Key Events 7.5 Distance 7.6 Convex Hulls 7.7 Implicit Paths Chapter 8 Discussion 8.1 Inquiry into Linguistic Universals 8.2 Falsifiability 8.3 The Nature of the Model 8.4 What Have We Learned about Models? 8.5 A Final Word Appendix Perceptual Primitives Center of Mass Major Axis Orientation Proximal and Center-of-Mass Orientations Learning in Map Comparison Structures

Terry Regier is Assistant Professor of Psychology at the University of Chicago.

등록된 사용후기

사용후기 쓰기 새 창 더보기

사용후기가 없습니다.

등록된 상품문의

상품문의 쓰기새 창 더보기

상품문의가 없습니다.

배송정보

교환/반품 정보