Conversational Recommender System

By Yueming Sun, Yi Zhang. Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, 2018

511 citations recorded by the source corpus.

A personalized conversational sales agent could have much commercial potential. E-commerce companies such as Amazon, eBay, JD, Alibaba etc. are piloting such kind of agents with their users. However, the research on this topic is very limited and existing solutions are either based on single round adhoc search engine or traditional multi round dialog system. They usually only utilize user inputs in the current session, ignoring users' long term preferences. On the other hand, it is well known that sales conversion rate can be greatly improved based on recommender systems, which learn user preferences based on past purchasing behavior and optimize business oriented metrics such as conversion rate or expected revenue. In this work, we propose to integrate research in dialog systems and recommender systems into a novel and unified deep reinforcement learning framework to build a personalized conversational recommendation agent that optimizes a per session based utility function. In particular, we propose to represent a user conversation history as a semi-structured user query with facet-value pairs. This query is generated and updated by belief tracker that analyzes natural language utterances of user at each step. We propose a set of machine actions tailored for recommendation agents and train a deep policy network to decide which action (i.e. asking for the value of a facet or making a recommendation) the agent should take at each step. We train a personalized recommendation model that uses both the user's past ratings and user query collected in the current conversational session when making rating predictions and generating recommendations. Such a conversational system often tries to collect user preferences by asking questions. Once enough user preference is collected, it makes personalized recommendations to the user. We perform both simulation experiments and real online user studies to demonstrate the effectiveness of the proposed framework.

Subfields: Artificial Intelligence, Information Systems

View Paper: https://www.semanticscholar.org/paper/255149101465960b3968ee00dd3f647ea4cf309a

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 Information Systems

Citations in the corpus

When large language models meet personalization: perspectives of challenges and opportunities (2023, unknown): “The attribute-based QA approach [24, 216–218] utilizes a pipeline method within the dialogue system.”
Towards Knowledge-Based Recommender Dialog System (2019, background): “, 2016), training the dialog agent to ask for facet values for recommendation (Sun and Zhang, 2018).”
Towards Deep Conversational Recommendations (2018, background): “Other approaches [15, 16, 17, 18] use reinforcement learning to train goal-oriented dialogue systems.”
Estimation-Action-Reflection: Towards Deep Interaction Between Conversational and Recommender Systems (2020, methodology): “ons, 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. CCS CONCE”
Improving Conversational Recommender Systems via Knowledge Graph based Semantic Fusion (2020, background): “a Mining (KDD ’20), August 23– 27, 2020, Virtual Event, USA. ACM, New York, NY, USA, 9 pages. https: //doi.org/10.1145/3394486.3403143 1 INTRODUCTION Recently, conversational recommender system (CRS) [4, 18, 19, 26, 34] has become an emerging research topic in seeking to provide high-quality recomme”
Advances and Challenges in Conversational Recommender Systems: A Survey (2021, methodology): “Zhang et al. [250] transform each textual review of part of the Amazon dataset into a question-answer sequence to simulate the conversation.”
A Survey of Graph Neural Networks for Recommender Systems: Challenges, Methods, and Directions (2021, background): “To address it, recently, conversational (interactive) recommendation researches [89, 115, 162, 204] proposed a new paradigm in which users interact with the information system through conversations, and then new data can be dynamically collected.”
Recommendation as Language Processing (RLP): A Unified Pretrain, Personalized Prompt & Predict Paradigm (P5) (2022, background): “The works mostly address four lines of research: 1) explainable recommendation [8, 25–27, 68, 70] where NLP models help generating text explanations for a given recommendation; 2) sequential recommendation as language modeling [7, 53, 73] which considers user interaction histories as word token sequ”
CRSLab: An Open-Source Toolkit for Building Conversational Recommender System (2021, methodology): “Besides, CRSLab also provides extensive automatic evaluation protocols and a human-machine interactive interface to compare and test different CRSs.”
RecSim: A Configurable Simulation Platform for Recommender Systems (2019, background): “…Christakopoulou et al., 2016], bandit-based exploration [Li et al., 2010, 2016, Christakopoulou and Banerjee, 2018], or reinforcement learning [Sun and Zhang, 2018] to explore the space of options in collaboration with the user to uncover good outcomes or maximize user engagement over…”
Continuous and Interactive Factual Knowledge Learning in Veriﬁcation Dialogues (2020, background): “Recently, many conversational systems have been proposed, e.g., Conversational Recommender Systems [46, 1, 58] and Information-seeking Conversation [55, 40].”
Model-Agnostic Counterfactual Reasoning for Eliminating Popularity Bias in Recommender System (2020, background): “Personalized recommendation has revolutionized a myriad of online applications such as e-commerce [48, 50], search engines [38], and conversational systems [26, 40].”
Memory Grounded Conversational Reasoning (2019, background): “(Bordes et al., 2017; Sun and Zhang, 2018), information querying (Williams et al.”
A Survey on Conversational Recommender Systems (2020, methodology): “the item embedding was based on the item’s textual description and the user’s query embedding encoded the user’s initial request and the follow-up conversations during the interaction. The authors of [133] also used a hybrid model, which used Factorization Machines to combine the dialogue state—repr”
OpenDialKG: Explainable Conversational Reasoning with Attention-based Walks over Knowledge Graphs (2019, background): “Most work focuses on task or goal oriented dialog systems such as conversational recommendations (Salem et al., 2014; Bordes et al., 2017; Sun and Zhang, 2018; Dalton, 2018), information querying (Williams et al., 2017; de Vries et al., 2018; Reddy et al., 2018), etc., with datasets collected…”
Reinforcement Learning based Recommender Systems: A Survey (2021, methodology): “[25] adapts REINFORCE algorithm to a neural candidate generator with a very large action space.”
Guided Transformer: Leveraging Multiple External Sources for Representation Learning in Conversational Search (2020, background): “the system asks about different attributes of the items to reveal the user preferences. For instance, Christakopoulou et al. [8] designed an interactive system for venue recommendation. Sun and Zhang [48] utilized facet-value pairs to represent a conversation history for conversational recommendatio”
How Can Recommender Systems Benefit from Large Language Models: A Survey (2023, unknown): “It can even provide ﬁne-grained control over different stages for recommendation ( e.g. , matching, ranking, reranking) Next, we will elaborate on the adaptation of LLM to different parts of the recommendation pipeline, except for user data collection.”
Asking Clarifying Questions in Open-Domain Information-Seeking Conversations (2019, unknown): “Also, Sun and Zhang [40] utilized a semi-structured user query with facet-value pairs to represent a conversation history and proposed a deep reinforcement learning framework to build a personalized conversational recommender system.”