Computational models of referring : a study in cognitive science /
Saved in:
| Author / Creator: | Deemter, Kees van, author. |
|---|---|
| Imprint: | Cambridge, Massachusetts : The MIT Press, [2016] |
| Description: | x, 339 pages : illustrations ; 24 cm |
| Language: | English |
| Subject: | |
| Format: | Print Book |
| URL for this record: | http://pi.lib.uchicago.edu/1001/cat/bib/11003784 |
Table of Contents:
- Preface
- I. First Part: Setting the Stage
- 1. Aims and Scope of This Book
- 1.1. Aims and Main Thesis
- 1.2. Reference in Practical Applications of Computing
- 1.3. Computational Models of Reference Production
- 1.4. Determining the Information Content of an RE
- 1.5. Focus on Speakers or Hearers?
- 1.6. Referring in One Shot
- 1.7. A Perspective on Reference: Information Sharing
- 1.8. Summary of the Chapter
- 2. Theories of Reference
- 2.1. What Makes a Referring Expression?
- 2.2. Knowing What Something Is
- 2.3. Denotation and Connotation
- 2.4. The Russell-Strawson Debate
- 2.5. Intensional Contexts
- 2.6. Attributive Descriptions and Misdescriptions
- 2.7. Proper Names
- 2.8. The Gricean Maxims and Relevance Theory
- 2.9. Summary of the Chapter
- 3. The Psychology of Reference Production
- 3.1. Common Ground
- 3.2. Audience Design and the Egocentricty Debate
- 3.3. Rationality and the Gricean Maxims
- 3.4. Intrinsic Preference for Certain Attributes
- 3.5. Comparing Preference with Discrimination
- 3.6. Insights from Dialogue
- 3.7. Ecological Validity of Experiments
- 3.8. Summary of the Chapter
- II. Second Part: Solving the Classic Reg Problem
- 4. Getting Computers to Refer
- 4.1. Computational Pre-history of REG
- 4.2. The California School
- 4.3. The Classic REG Task
- 4.4. Assumptions Behind the Classic REG Task
- 4.5. Exploring the Gricean Angle Computationally
- 4.6. The Incremental Algorithm
- 4.7. Logical (In)completeness
- 4.8. Computational Tractability of REG Algorithms
- 4.9. Salience
- 4.10. Summary of the Chapter
- 5. Testing REG Algorithms: The TUNA Experiment
- 5.1. Why the TUNA Experiment?
- 5.2. How to Test a REG Algorithm?
- 5.3. The TUNA Corpus and Its Annotation
- 5.4. Analysis of the Furniture Corpus
- 5.5. Analysis of the People Corpus
- 5.6. Modelling a Plurality of Speakers
- 5.7. Lessons from the TUNA Experiment
- 5.8. Lessons from the TUNA Evaluation Challenges
- 5.9. A Note on Alternative Metrics
- 5.10. Summary of the Chapter
- 6. Probabilistic and Other Alternatives to the Classic REG Algorithms
- 6.1. Variations in Language Production
- 6.2. Bayesian Models of Reference
- 6.3. Probabilistic Referential Overspecification: the PRO Algorithm
- 6.4. Constraint Satisfaction for REG
- 6.5. Krahmer et al.'s Cost-Based Approach
- 6.6. Appek's Heirs: Reference as Part of a Wider Problem
- 6.7. Summary of the Chapter
- III. Third Part: Generating a Wider Class of Res
- 7. First Extension: Using Proper Names
- 7.1. Why Have Proper Names Been Neglected in REG?
- 7.2. Incorporating Proper Names into REG
- 7.3. Reifying Properties
- 7.4. Challenges for REG Posed by Proper Names
- 7.5. Summary of the Chapter
- 8. Second Extension: Referring to Sets
- 8.1. Purely Conjunctive References to Sets
- 8.2. Negation and Disjunction
- 8.3. Satellite Sets and Their Use in REG
- 8.4. Generating Boolean Logical Forms Incrementally
- 8.5. Optimization of Generated REs
- 8.6. Issues Raided by the Algorithms Proposed
- 8.7. Lexical Coherence in Conjoined RES
- 8.8. Avoiding Surface Ambiguities
- 8.9. Beyond Sets of Objects
- 8.10. Summary of the Chapter
- 9. Third Extension: Using Gradable Properties
- 9.1. The Semantics of Vague Descriptions
- 9.2. Pragmatic Constraints on What Can Be Said
- 9.3. Empirical Grounding
- 9.4. Computational Generation of Vague Descriptions
- 9.5. Puzzles for Incremental Content Determination
- 9.6. A Case Study: Real-World Objects and Their Sizes
- 9.7. Can We Ever Be Clear? Salience as a Gradable Property
- 9.8. Summary of the Chapter
- 10. Fourth Extension: Exploiting Modern Knowledge Representation
- 10.1. Knowledge Representation and REG
- 10.2. Description Logic: a Primer
- 10.3. Applying Description Logic to Familiar REG Problems
- 10.4. Exploiting the Full Power of Di.
- 10.5. Using SROTQ + to Generate Complex REs
- 10.6. Rethinking REG: Using Shared Knowledge That Is Not Atomic
- 10.7. Why Study the Generation of Logically Complex Res?
- 10.8. Summary of the Chapter
- 11. The Question of Referability
- 11.1. Revisiting the Logical Completeness of REG
- 11.2. Limitations of SROIQ + and the GROWL Algorithm
- 11.3. Even More Expressive Algorithms?
- 11.4. Summary of the Chapter
- IV. Fourth Part: Generalizing Reference Generation
- 12. First Challenge: Large Domains
- 13. Second Challenge: Breakdown of Common Knowledge
- 14. Third Challenge: Approximate Reference
- 15. Fourth Challenge: Going Beyond Identification
- Summary of Part IV: Complexities of Information Sharing
- V. Epilogue
- 16. Epilogue
- 16.1. REG Algorithms as Cognitive Models
- 16.2. The Gricean Maxims and the Principle of Intrinsic Preference
- 16.3. Future Research: The Way Ahead
- Frequently Occurring Terms and Abbreviations
- Bibliography
- Index
