Leveraging Historical Interaction Data for Improving Conversational Recommender System
This work addresses conversational recommender systems, which are an emerging practical topic, by incorporating historical data to potentially enhance recommendation accuracy.
The authors tackled the problem of conversational recommender systems by leveraging historical interaction data alongside conversation data, proposing a novel pre-training approach that integrates item- and attribute-based preferences with two pre-training tasks and a negative sample generator. Their method demonstrated effectiveness on two real-world datasets.
Recently, conversational recommender system (CRS) has become an emerging and practical research topic. Most of the existing CRS methods focus on learning effective preference representations for users from conversation data alone. While, we take a new perspective to leverage historical interaction data for improving CRS. For this purpose, we propose a novel pre-training approach to integrating both item-based preference sequence (from historical interaction data) and attribute-based preference sequence (from conversation data) via pre-training methods. We carefully design two pre-training tasks to enhance information fusion between item- and attribute-based preference. To improve the learning performance, we further develop an effective negative sample generator which can produce high-quality negative samples. Experiment results on two real-world datasets have demonstrated the effectiveness of our approach for improving CRS.