New opening for an Internship position: Comparing Topic Models and Concept Networks for measuring text similarity

The goal of the project is to compare two types of advanced methods for semantic similarity that make use of knowledge about relationships between words, learned from an initial set of documents. These methods are developed by the supervisors' research teams. The first method uses probabilities of word co-occurrences to infer "semantic topics", which are synonym-like relationships. The second method makes use of Wikipedia to determine conceptual relations between words or sets of words.

This is a Master internship position for EPFL students only.

*** This position has been filled ***

 

More information about available positions can be found on our online recruitment system