A Survey on Conversational Recommender Systems

Jannach, Dietmar; Manzoor, Ahtsham; Cai, Wanling; Chen, Li

doi:10.48550/arxiv.2004.00646

Cited by 17 publications

(37 citation statements)

References 112 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Besides, our method aims to ground the semantic meaning of the comparison view into the dimensions of disentangled representations. Critiquing Recommender Systems Critiquing is a method widely used conversational recommendation [30] which supports a taskoriented, multi-turn dialogue with their users to discover the detailed and current preferences of the user [14]. In critiquing approaches, users are presented with a recommendation result during the dialogue and then apply pre-defined critiques on the result [4,10].…”

Section: Related Work Product Search and Item Retrievalmentioning

confidence: 99%

Controllable Gradient Item Retrieval

Wang

Zhou

Yang

et al. 2021

Proceedings of the Web Conference 2021

View full text Add to dashboard Cite

In this paper, we identify and study an important problem of gradient item retrieval. We define the problem as retrieving a sequence of items with a gradual change on a certain attribute, given a reference item and a modification text. For example, after a customer saw a white dress, she/he wants to buy a similar one but more floral on it. The extent of "more floral" is subjective, thus prompting one floral dress is hard to satisfy the customer's needs. A better way is to present a sequence of products with increasingly floral attributes based on the white dress, and allow the customer to select the most satisfactory one from the sequence. Existing item retrieval methods mainly focus on whether the target items appear at the top of the retrieved sequence, but ignore the demand for retrieving a sequence of products with gradual change on a certain attribute. To deal with this problem, we propose a weakly-supervised method that can learn a disentangled item representation from user-item interaction data and ground the semantic meaning of attributes to dimensions of the item representation. Our method takes a reference item and a modification as a query. During inference, we start from the reference item and "walk" along the direction of the modification in the item representation space to retrieve a sequence of items in a gradient manner. We demonstrate our proposed method can achieve disentanglement through weak supervision. Besides, we empirically show that an item sequence retrieved by our method is gradually changed on an indicated attribute and, in the item retrieval task, our method outperforms existing approaches on three different datasets.

show abstract

Section: Related Work Product Search and Item Retrievalmentioning

confidence: 99%

Controllable Gradient Item Retrieval

Wang

Zhou

Yang

et al. 2021

Proceedings of the Web Conference 2021

View full text Add to dashboard Cite

show abstract

“…Moreover, user's demands may also vary over time. Thus, it is very hard for traditional recommendation systems to capture user's dynamic interests timely [10].…”

Section: Introductionmentioning

confidence: 99%

“…A CRS can interact with users by natural language and obtain users' explicit feedback timely for better understanding users' dynamic interests. Previous studies about CRS can be classified into two main categories: 1) attribute-based CRS, and 2) chit-chat-based CRS [10]. The attribute-based conversational recommendation methods [3,4,15,16,29,35,40] mainly focus on the recommendation task.…”

Section: Introductionmentioning

confidence: 99%

“…The effect of the KG in the response generation module may also be limited. Moreover, KG entities' embeddings are learned in the recommendation module while words embeddings are learned in the generation module [2,10,38]. The representation space gap between these two embeddings may also limit the capability of the models to generate KG entities in responses.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

KECRS: Towards Knowledge-Enriched Conversational Recommendation System

Zhang¹,

Liu²,

Zhong³

et al. 2021

Preprint

View full text Add to dashboard Cite

The chit-chat-based conversational recommendation systems (CRS) provide item recommendations to users through natural language interactions. To better understand user's intentions, external knowledge graphs (KG) have been introduced into chit-chat-based CRS. However, existing chit-chat-based CRS usually generate repetitive item recommendations, and they cannot properly infuse knowledge from KG into CRS to generate informative responses. To remedy these issues, we first reformulate the conversational recommendation task to highlight that the recommended items should be new and possibly interested by users. Then, we propose the Knowledge-Enriched Conversational Recommendation System (KECRS). Specifically, we develop the Bag-of-Entity (BOE) loss and the infusion loss to better integrate KG with CRS for generating more diverse and informative responses. BOE loss provides an additional supervision signal to guide CRS to learn from both human-written utterances and KG. Infusion loss bridges the gap between the word embeddings and entity embeddings by minimizing distances of the same words in these two embeddings. Moreover, we facilitate our study by constructing a high-quality KG, i.e., The Movie Domain Knowledge Graph (TMDKG). Experimental results on a large-scale dataset demonstrate that KECRS outperforms state-of-the-art chitchat-based CRS, in terms of both recommendation accuracy and response generation quality.

show abstract

“…However, these methods may fail to model short term or dynamic preferences of a user. Subsequently, methods to tackle this explicitly problem have been proposed, for example, through conversations via a conversational recommender system [13], or critiquing recommendations using language or keyphrases [18,23,35,37]. In contrast to keyphrases, typically mined from reviews or descriptions (for instance), this work considers building controllable or critique-able recommenders using attribute data for items, which can be used to construct preference distributions at a user level.…”

Section: Introductionmentioning

confidence: 99%

Controllable Recommenders using Deep Generative Models and Disentanglement

Bhargav,

Kanoulas

2021

Preprint

View full text Add to dashboard Cite

In this paper, we consider controllability as a means to satisfy dynamic preferences of users, enabling them to control recommendations such that their current preference is met. While deep models have shown improved performance for collaborative filtering, they are generally not amenable to fine grained control by a user, leading to the development of methods like deep language critiquing. We propose an alternate view, where instead of keyphrase based critiques, a user is provided 'knobs' in a disentangled latent space, with each knob corresponding to an item aspect. Disentanglement here refers to a latent space where generative factors (here, a preference towards an item category like genre) are captured independently in their respective dimensions, thereby enabling predictable manipulations, otherwise not possible in an entangled space. We propose using a (semi-)supervised disentanglement objective for this purpose, as well as multiple metrics to evaluate the controllability and the degree of personalization of controlled recommendations. We show that by updating the disentangled latent space based on user feedback, and by exploiting the generative nature of the recommender, controlled and personalized recommendations can be produced. Through experiments on two widely used collaborative filtering datasets, we demonstrate that a controllable recommender can be trained with a slight reduction in recommender performance, provided enough supervision is provided. The recommendations produced by these models appear to both conform to a user's current preference and remain personalized.

show abstract

A Survey on Conversational Recommender Systems

Cited by 17 publications

References 112 publications

Controllable Gradient Item Retrieval

Controllable Gradient Item Retrieval

KECRS: Towards Knowledge-Enriched Conversational Recommendation System

Controllable Recommenders using Deep Generative Models and Disentanglement

Contact Info

Product

Resources

About