The goal is to provide a broad overview of processing natural language texts, levels of representation of natural language, and representations of knowledge about natural language.

Areas and classical tasks and applications of Natural Language Processing (NLP):

• Language models, evaluation methods, and evaluation metrics
• Word level: Edit Distance, word representations (WordNet, word vectors)
• Sentence level: Named Entity Recognition, Part-Of-Speech tagging, parsing
• Text level: Text Classification, Sentiment Analysis, Distributional Semantics
• Explicit and implicit knowledge; syntactic, semantic, and world knowledge
• Applications: Search Engines, Question Answering, Knowledge Mining, Turing test

Exercises

• Supervised machine learning exercise (sequence tagging)
• Unsupervised machine learning exercise (distributional semantics)

ECTS Breakdown: VU 2.0 h, 3 ECTS = 75 hours

• Lecture 18 h
• Reading 25 h
• Discussion of Exercises 2 x 1 h = 2 h
• Solving Exercises 2 x 15 h = 30 h

Literature:

• Daniel Jurafsky and James Martin: Speech and Language Processing: An Introduction to Natural Language Processing, Computational Linguistics, and Speech Recognition. Prentice Hall, 2000.

Attendance required!

• Active participation during the lecture, in particular discussions about reading assignments
• Project exercises (programming, evaluation, writing project report)
• Oral exam (if required).

Due to the high demand for this course, in the introductory lesson a placement exercise will be distributed that has to be handed in per email and will determine who will get one of the 30 places in the course. This exercise has the purpose of finding out who is most interested in the course. The order of registration is not relevant.

Basic Python knowledge will be an advantage.