Estimation-Action-Reflection: Towards Deep Interaction Between Conversational and Recommender Systems

By Wenqiang Lei, Xiangnan He, Yisong Miao, Qingyun Wu, Richang Hong, Min-Yen Kan, Tat-seng Chua. Web Search and Data Mining, 2020

340 citations recorded by the source corpus.

Recommender systems are embracing conversational technologies to obtain user preferences dynamically, and to overcome inherent limitations of their static models. A successful Conversational Recommender System (CRS) requires proper handling of interactions between conversation and recommendation. We argue that three fundamental problems need to be solved: 1) what questions to ask regarding item attributes, 2) when to recommend items, and 3) how to adapt to the users' online feedback. To the best of our knowledge, there lacks a unified framework that addresses these problems. In this work, we fill this missing interaction framework gap by proposing a new CRS framework named Estimation"Action" Reflection, or EAR, which consists of three stages to better converse with users. (1) Estimation, which builds predictive models to estimate user preference on both items and item attributes; (2) Action, which learns a dialogue policy to determine whether to ask attributes or recommend items, based on Estimation stage and conversation history; and (3) Reflection, which updates the recommender model when a user rejects the recommendations made by the Action stage. We present two conversation scenarios on binary and enumerated questions, and conduct extensive experiments on two datasets from Yelp and LastFM, for each scenario, respectively. Our experiments demonstrate significant improvements over the state-of-the-art method CRM [32], corresponding to fewer conversation turns and a higher level of recommendation hits.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Information Systems

View Paper: https://www.semanticscholar.org/paper/1abe8ff4bd56c93b3c6c560783e1949d104546e9

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Information Systems

Citations in the corpus

Parameter-Efficient Transfer from Sequential Behaviors for User Modeling and Recommendation (2020, methodology): “As for the fine-tuning objective functions of PeterRec, we adopt the pairwise ranking loss (BPR) [20, 28] for top-N item recommendation task and the CE loss for the user profile classification tasks.”
CRSLab: An Open-Source Toolkit for Building Conversational Recommender System (2021, methodology): “For the conversation task, CRSLab supports both relevance-based (i.e., BLEU (Papineni et al., 2002) and embedding-based metrics (Liu et al., 2016)) and diversity-based evaluation metrics (i.e., Distinct-{1,2,3,4} (Li et al., 2016)).”
Disentangled Graph Collaborative Filtering (2020, background): “history with users [21], item knowledge [15, 32, 39, 41, 42], and user reviews [5, 22], or conduct psychology experiments [24] to establish the ground truth on user intents and further interpret disentangled representations better.”
Recommendation as Instruction Following: A Large Language Model Empowered Recommendation Approach (2023, neutral): “Specially, LLMs act as a traditional retriever to process users’ vague or specific queries [20].”
When large language models meet personalization: perspectives of challenges and opportunities (2023, neutral): “The attribute-based QA approach [185], [198], [199], [200] utilizes a pipeline”
Improving Conversational Recommender Systems via Knowledge Graph based Semantic Fusion (2020, background): “Recently, several studies started to integrate the two components for understanding users’ needs and recommend the right items through natural language conversation [16, 26, 35].”