Speaker Role Contextual Modeling for Language Understanding and Dialogue Policy Learning
This work addresses the challenge of handling random and goal-driven dialogues in conversational systems, though it appears incremental as it builds on existing contextual modeling approaches.
The paper tackled the problem of modeling unpredictable human-human dialogues by proposing a role-based contextual model that independently considers different speaker roles based on their speaking patterns. The result was a significant improvement in language understanding and dialogue policy learning tasks on a benchmark dataset.
Language understanding (LU) and dialogue policy learning are two essential components in conversational systems. Human-human dialogues are not well-controlled and often random and unpredictable due to their own goals and speaking habits. This paper proposes a role-based contextual model to consider different speaker roles independently based on the various speaking patterns in the multi-turn dialogues. The experiments on the benchmark dataset show that the proposed role-based model successfully learns role-specific behavioral patterns for contextual encoding and then significantly improves language understanding and dialogue policy learning tasks.