Language Learning 2007: During the course the site will be updated on a weekly basis

Schedule Information

Lecturer: Prof. dr. Pieter Adriaans (email: pietera@science.uva.nl).
Practice sessions: Sophia Katrenko (email: katrenko@science.uva.nl).
Level: III
UvA credits: 6
Semester 2 (first and second period) (5 February - 21 May)
Classes:   Mondays, 15.00-17.00, Room P-0.17 (lectures)
There will be no separate practice classes. The assignments will be handed out during the course.
Examinations: see below

Course Description

In this class you will learn the most important concepts in the field of language learning. Last year language learning as data compression and the minimum description principle got special attention. This year we will focus on questions concerning mathematical models of the learning capacity of human beings. This leads to the study of syntactical learning algorithms. We will explore the possibilities to extend this theory to a general theory of learnable structures. Topics: Identification in the limit, Locking sequences, Finite Elasticity, Minimum Description Length, Randomness deficiency, Normalized compression distance Formalisms, DEC grammars, Regular Languages, DFA, Context free grammars, Algorithms, EDSM, EMILE, ADIOS. Semantic Learning, Real Life corpora, The complexity of natural language, Text Mining.

Background Knowledge

Familiarity with the theory of Languages and Automata as treated in Hopcroft and Ullmann. Experience with one or more programming languages.

Examinations

This course will be graded based on assignments that have to be handed in each week. There are two kinds of assignemtns: individual assignments and programming assignments. The individual assignments consist of questions that you have to answer in writing yourself. The programming assignments can be made in any programming language. For the programming assignments you can work together in groups of maximum three people. We suggest that you take a language that you already master well. The Exel sheet with the marks is here.

The assignments will be made available here at least two days before the practice session at which they will be handed out. Individual assignments have to be handed in weekly, no later than Monday (09:00). It is allowed to use an extra-week for the programming assignments (again, please submit them no later than Monday (09:00). Late submissions will be penalized. Please send your answers (code, executables, etc.) to Sophia Katrenko with a subject line "LL07 (homework)".

Online Course Materials

Weekly slides and assignments

Background Material

Learnability and identification in the limit
Grammars & grammar formalisms
MDL & Kolmogorov complexity
Grammatical inference
Neurocomputational processing

Offsite links


Sophia Katrenko

Last modified: Jan 2006