A Deep Reinforcement Learning Approach for Interactive Search with Sentence-level Feedback
This addresses the challenge of incorporating sentence-level feedback in interactive search systems, which is incremental but improves accuracy for users.
The paper tackles the problem of interactive search systems that ignore fine-grained sentence-level feedback by proposing DQrank, a deep Q-learning approach that adapts BERT models to select crucial sentences and rank items, achieving at least 12% better performance than previous state-of-the-art RL methods on three datasets.
Interactive search can provide a better experience by incorporating interaction feedback from the users. This can significantly improve search accuracy as it helps avoid irrelevant information and captures the users' search intents. Existing state-of-the-art (SOTA) systems use reinforcement learning (RL) models to incorporate the interactions but focus on item-level feedback, ignoring the fine-grained information found in sentence-level feedback. Yet such feedback requires extensive RL action space exploration and large amounts of annotated data. This work addresses these challenges by proposing a new deep Q-learning (DQ) approach, DQrank. DQrank adapts BERT-based models, the SOTA in natural language processing, to select crucial sentences based on users' engagement and rank the items to obtain more satisfactory responses. We also propose two mechanisms to better explore optimal actions. DQrank further utilizes the experience replay mechanism in DQ to store the feedback sentences to obtain a better initial ranking performance. We validate the effectiveness of DQrank on three search datasets. The results show that DQRank performs at least 12% better than the previous SOTA RL approaches. We also conduct detailed ablation studies. The ablation results demonstrate that each model component can efficiently extract and accumulate long-term engagement effects from the users' sentence-level feedback. This structure offers new technologies with promised performance to construct a search system with sentence-level interaction.