Intent Classification by the use of Automatically Generated Knowledge Graphs

Anonymous

Abstract: Intent classification is an essential task for goal-oriented dialogue systems, in order to automatically identify customers' goals. Although intent classification performs well in general settings, domain-specific user goals can still present a challenge for this task. To address this challenge, we automatically generate knowledge graphs for targeted datasets to capture domain-specific knowledge and leverage embeddings trained on these knowledge graphs for the intent classification task. We compare our results with state-of-the-art pre-trained sentence embeddings. Our evaluation on three datasets show improvement on the intent classification task in terms of precision.

Paper Type: long