INSPIRED: Toward Sociable Recommendation Dialog Systems

Hayati, Shirley Anugrah; Kang, Dongyeop; Zhu, Qingxiaoyang; Shi, Weiyan; Yu, Zhou

doi:10.18653/v1/2020.emnlp-main.654

Cited by 58 publications

(35 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Open-ended CRS. Recently, researchers begin to explore the more free-style item recommendation in the response generation, i.e., open-ended CRS Chen et al, 2019;Liao et al, 2019;Kang et al, 2019;Zhou et al, 2020a;Ma et al, 2020;Hayati et al, 2020;Zhou et al, 2020b;Zhang et al, 2021). Generally, this kind of systems consist of two major components, namely a recommender component to recommend items and a dialogue component to generate natural responses.…”

Section: Related Workmentioning

confidence: 99%

Learning Neural Templates for Recommender Dialogue System

Liang¹,

Hu²,

Xu³

et al. 2021

Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing

View full text Add to dashboard Cite

Though recent end-to-end neural models have shown the promising progress on Conversational Recommender System (CRS), two key challenges still remain. First, the recommended items cannot be always incorporated into the generated replies precisely and appropriately. Second, only the items mentioned in the training corpus have a chance to be recommended in the conversation. To tackle these challenges, we introduce a novel framework called NTRD for recommender dialogue system that decouples the dialogue generation from the item recommendation. NTRD has two key components, i.e., response template generator and item selector. The former adopts an encoder-decoder model to generate a response template with slot locations tied to target items, while the latter fills in slot locations with the proper items using a sufficient attention mechanism. Our approach combines the strengths of both classical slot filling approaches (that are generally controllable) and modern neural NLG approaches (that are generally more natural and accurate). Extensive experiments on the benchmark RE-DIAL show our NTRD significantly outperforms the previous state-of-the-art methods. Besides, our approach has the unique advantage to produce novel items that do not appear in the training set of dialogue corpus. The code is available at https://github. com/jokieleung/NTRD. * Equal contribution. Work performed when Zujie Liang was an intern at Microsoft STCA.

show abstract

Section: Related Workmentioning

confidence: 99%

Learning Neural Templates for Recommender Dialogue System

Liang¹,

Hu²,

Xu³

et al. 2021

Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing

View full text Add to dashboard Cite

show abstract

“…Mainly, in these approaches, multiple neural network components including RNNs, GCNs, CNNs, GANs, encoder-decoders pairs etc. were used to carry out the task of language generation and recommendation [6,10,11,12,13,14,21]. For example, in the work of Li et al [10], an RNN module is responsible for predicting the sentiment of user's utterance towards a given a preference (entity), and subsequently the outcomes of the RNN module are used as an input to the autonencoder-based module for computing the recommendation.…”

Section: Language Generation Approaches: Kbrd and Kgsfmentioning

confidence: 99%

“…In total, there were 34 unique dialog situations. 14 The average scores and standard deviations for both unique and total cases are shown in Table 5.…”

Section: Analysis Of Failure Situations: Specific Seeker Queriesmentioning

confidence: 99%

“…In recent years, different deep neural network architectures were explored for this language generation task with the aim to continuously improve the quality of the generated responses. Such architectures for example include recurrent neural networks (RNN), sequenceto-sequence transformer and encoder-decoder models [10,11,12,13,14]. Furthermore, these networks are sometimes coupled with additional components and external knowledge sources for the recommendation task, e.g., in [6,11].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Towards Retrieval-based Conversational Recommendation

Manzoor,

Jannach

2021

Preprint

View full text Add to dashboard Cite

Conversational recommender systems have attracted immense attention recently. The most recent approaches rely on neural models trained on recorded dialogs between humans, implementing an end-to-end learning process. These systems are commonly designed to generate responses given the user's utterances in natural language. One main challenge is that these generated responses both have to be appropriate for the given dialog context and must be grammatically and semantically correct. An alternative to such generation-based approaches is to retrieve responses from pre-recorded dialog data and to adapt them if needed. Such retrieval-based approaches were successfully explored in the context of general conversational systems, but have received limited attention in recent years for CRS. In this work, we re-assess the potential of such approaches and design and evaluate a novel technique for response retrieval and ranking. A user study (N=90) revealed that the responses by our system were on average of higher quality than those of two recent generation-based systems. We furthermore found that the quality ranking of the two generation-based approaches is not aligned with the results from the literature, which points to open methodological questions. Overall, our research underlines that retrieval-based approaches should be considered an alternative or complement to language generation approaches. 1

show abstract

“…Multiple paths in KG help to locate the subspace of the user's interest and generate interpretive utterances in line with people's dialogue behavior, e.g., "...Inception with Christopher Nolan and Leonardo DiCaprio...". However, KGs can not record all relations of interest entities involved in real-world diverse dialogues (Ma et al, 2020;Hayati et al, 2020;Sarkar et al, 2020). Therefore, it is often difficult to achieve a complete reasoning path of interest shift within limited number of hops.…”

Section: Introductionmentioning

confidence: 99%

CRFR: Improving Conversational Recommender Systems via Flexible Fragments Reasoning on Knowledge Graphs

Zhou¹,

Wang²,

He³

et al. 2021

Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing

View full text Add to dashboard Cite

Although paths of user interests shift in knowledge graphs (KGs) can benefit conversational recommender systems (CRS), explicit reasoning on KGs has not been well considered in CRS, due to the complex of high-order and incomplete paths. We propose CRFR, which effectively does explicit multi-hop reasoning on KGs with a conversational context-based reinforcement learning model. Considering the incompleteness of KGs, instead of learning single complete reasoning path, CRFR flexibly learns multiple reasoning fragments which are likely contained in the complete paths of interests shift. A fragments-aware unified model is then designed to fuse the fragments information from item-oriented and concept-oriented KGs to enhance the CRS response with entities and words from the fragments. Extensive experiments demonstrate CRFR's SOTA performance on recommendation, conversation and conversation interpretability.

show abstract

INSPIRED: Toward Sociable Recommendation Dialog Systems

Cited by 58 publications

References 27 publications

Learning Neural Templates for Recommender Dialogue System

Learning Neural Templates for Recommender Dialogue System

Towards Retrieval-based Conversational Recommendation

CRFR: Improving Conversational Recommender Systems via Flexible Fragments Reasoning on Knowledge Graphs

Contact Info

Product

Resources

About