Named Entities for Computational Linguistics

Damien Nouvel is Associate Professor at the National Institute of Oriental Languages And Civilizations (Inalto) in Paris, France. Maud Ehrmann is a Research Scientist at EPFL (École polytechnique fédérale de Lausanne) in Geneva, Switzerland. Sophie Rosset is a Senior Researcher at the French National Centre for Scientific Research (CNRS) in Paris, France.

• Introduction  ixChapter 1. Named Entities for Accessing Information  11.1. Research program history 21.1.1. Understanding documents: an ambitious task 21.1.2. Detecting basic elements: named entities  31.1.3. Trend: a return to slot filling 71.2. Task using named entities as a basic representation 91.3. Conclusion 10Chapter 2. Named Entities, Referential Units 112.1. Issues with the named entity concept 122.1.1. A heterogeneous set 122.1.2. Existing defining formulas 172.1.3. An NLP object 212.2. The notions of meaning and reference 222.2.1. What is the reference? 222.2.2. What is meaning? 242.3. Proper names 272.3.1. The traditional criteria for defining a proper name 282.3.2. Meaning and referential function of proper names 302.3.3. The “referential load” of proper names 342.4. Definite descriptions 352.4.1. What is a definite description? 352.4.2. The meaning of definite descriptions 382.4.3. Complete and incomplete definite descriptions 392.5. The meaning and referential functioning of named entities  412.5.1. Reference to a particular 422.5.2. Referential autonomy 442.5.3. A “natural” heterogeneity 452.6. Conclusion 46Chapter 3. Resources Associated with Named Entities 473.1. Typologies: general and specialist domains 483.1.1. The notion of category 483.1.2. Typology development 493.1.3. Typologies beyond evaluation campaigns 533.1.4. Other uses of typologies 543.1.5. Illustrated comparison 573.1.6. Issues to consider regarding entities 573.2. Corpora 593.2.1. Introduction . 593.2.2. Corpora and named entities 603.2.3. Conclusion 653.3. Lexicons and knowledge databases 653.3.1. Lexical databases 663.3.2. Knowledge databases 723.4. Conclusion 75Chapter 4. Recognizing Named Entities 774.1. Detection and classification of named entities 784.2. Indicators for named entity recognition 794.2.1. Describing word morphology 794.2.2. Using lexical databases 814.2.3. Contextual clues 834.2.4. Conclusion 854.3. Rule-based techniques 854.4. Data-driven and machine-learning systems 884.4.1. Majority class models 914.4.2. Contextual models (HMM) 924.4.3. Multiple feature models (Softmax and MaxEnt) 934.4.4. Conditional Random Fields (CRFs) 954.5. Unsupervised enrichment of supervised methods 954.6. Conclusion 96Chapter 5. Linking Named Entities to References 995.1. Knowledge bases 1005.2. Formalizing polysemy in named entity mentions 1025.3. Stages in the named entity linking process 1035.3.1. Detecting mentions of named entities 1035.3.2. Selecting candidates for each mention 1035.3.3. Entity disambiguation 1045.3.4. Entity linking 1065.4. System performance 1065.4.1. Practical application: DBpedia Spotlight 1075.4.2. Future prospects 108Chapter 6. Evaluating Named Entity Recognition 1116.1. Classic measurements: precision, recall and F-measures 1126.2. Measures using error counts 1156.3. Evaluating associated tasks 1206.3.1. Detecting entities and mentions 1216.3.2. Entity detection and linking 1226.4. Evaluating preprocessing technologies 1266.5. Conclusion 128Conclusion  131Appendices  137Appendix 1. Glossary 139Appendix 2. Named Entities: Research Programs 141Appendix 3. Summary of Available Corpora 147Appendix 4. Annotation Formats 151Appendix 5. Named Entities: Current Definitions 153Bibliography 157Index 169