Best results in TAC KBP Event Nugget Detection

Peter Makarov and Simon Clematide showed the best scores in the trilingual Event Nugget Detection Track of this year's TAC KBP Evaluation. Peter and Simon participated with a structured prediction RNN model in the task of identifying Event Nuggets (event mentions) in the text, their types and factuality. Their system demonstrated the best results in all three languages. TAC (Text Analysis Conference) is a leading platform for information extraction evaluations, a successor to ACE and MUC.

Link to the poster: