Information Extraction (TPT33 / INF393)

Content

In this class, we will take an overview of Information Extraction techniques and the Semantic Web. Information extraction is the process of deriving structured information (such as alive(Elvis)) from digital text (such as the sentence “Elvis is alive”). The first part of this lecture will focus on factual and semantic information extraction, i.e., we will cover named entity recognition, entity disambiguation, instance extraction, fact extraction, and ontological information extraction. The Semantic Web is the little brother of the Web that aims to represent information in a machine-readable form. So, after having learned how to extract the information from text documents, we will learn how to represent it in a semantic way. We will cover the standards RDF/S, URIs, and RDFa, and recent advances in the field. We will also touch upon applications of both Information Extraction and the Semantic Web, such as Google's knowledge graph, IBM's Watson question answering system, and Facebook's Open Graph, and academic projects such as YAGO, DBpedia, and NELL.

Organization

Course title: Information Extraction
Course id in ATHENS: TPT33
Course id at Télécom: INF393
URL: https://suchanek.name/work/teaching/athens-2017
Language: English
Location: Télécom ParisTech, 46 rue Barrault, 75013 Paris, France
Time: Monday 2017-11-13 to Friday 2017-11-17
Schedule: every day

The class will be evaluated by work in the labs. The labs will be a combination of practical work (programming) and exam-like exercises. Every student works on their own.

Schedule

Day Session 1Session 2LabLab room
Monday Intro, Motivation Knowledge Representation, Corpus, Character EncodingsKnowledge RepresentationC45
Tuesday Named Entity Recognition Evaluation, DisambiguationResearchDisambiguationC128
Wednesday Instance Extraction, Fact Extraction, Wrapper induction POS Tagging, Dependency Parsing Type extractionC128
Thursday IE by Reasoning, Markov Logic SecuritySecurityC128
Friday Rule Mining, Description Logics Semantic WebEntity LinkingC130
The PDF slides are provided for convenience only, the authoritative ones are the SVG slides.