The growing availability of language corpora combined with large computing power has opened up exciting new possibilities for computer-aided language processing. ILK is a research programme in which inductive and memory-based learning algorithms are used for solving natural language problems.

An important component of the research carried out by ILK is the application of the developed learning methods to specific problems in language technology. Currently, ILK is involved in the following three major application areas:

• Memory-based Learning Applications in Language Technology

  • for speech synthesis:
    • grapheme-phoneme conversion
    • sentence accent and phrase break placement
    • word stress placement
    • morphological analysis
    • syllabification, hyphenation
  • for text analysis and processing:
    • morpho-syntactic word-class tagging
    • constituent chunking
    • word sense disambiguation
    • structural disambiguation
    • subcategorisation
    • shallow parsing

• Information Retrieval

  • for searching in document data bases and automatic content extraction; unsupervised learning (clustering)
  • for reusable distributed lexical representations of words (Lexspace).

• Development of Data Mining techniques

  for exploring, visualising, and exploiting large data bases of classification tasks.

ILK - Current Projects

Machine Learning as a Tool in Linguistic Research

CLS PhD Project, Feb 1995 - Feb 1999.
Jakub Zavrel

Memory-Based Syntactic Analysis

CLS PhD Project, Jan 1997 - Dec 2000.
Jorn Veenstra

Machine Learning Techniques for Linguistic Aspects of Speech Synthesis

SoBU PhD Project, Jan 1997 - Dec 2000.
Bertjan Busser

Automatic Lexical Acquisition: Subcategorisation Frames

NWO PhD Project, May 1997 - April 2001.
Sabine Buchholz

Machine Learning Algorithms for Natural Language Learning

NWO Postdoc Project, from July 1999
Khalil Sima'an

Memory Models of Language

KNAW Research Fellowship, from July 1999
Antal van den Bosch

Automatic inductive learning techniques for planning systems

SoBU PhD Project, from September 1999
Laura Maruster

Scientific Programming

Ko van der Sloot

Last update: 24 January 2000