The UvT Expert Collection is a test collection for use in 'enterprise search' experiments, similar to the TREC W3C collection. It was harvested from the Webwijs system developed at Tilburg University.

The ILK test collection contains 147 papers in ASCII format (co-)authored by current or ex-ILK members, accompanied by 80 natural language queries and the corresponding relevance assignments. It also contains a list of the authors associated with each paper.

Compiled by Jörg Tiedemann (Rijksuniversiteit Groningen). Contains translation models computed with GIZA++, and bilingual alignments of the Europarl corpus compiled by Philipp Koehn.

A movie (MP4, 23 Mb) explaining the basics of k-Nearest Neighbor classification, applied to English word pronunciation. Won the Best Explanation award at the AAAI-07 AI Video Competition.