Posted: 27th June 2014

A paper by two Oxford computer scientists on modelling word structure received the award for the best application paper at this year's International Conference on Machine Learning (ICML). The paper by Jan Botha and Phil Blunsom is entitled Compositional Morphology for Word Representations and Language Modelling, and it introduces a simple but effective technique for integrating morphological information into statistical language models based on vector representations. The work was presented at ICML, which took place in Beijing on 22-24 June.

The paper and slides are available online, and source code will be available soon.