Multi-Behavior Sequential Recommendation with Temporal Graph Transformer

Xia, Lianghao; Huang, Chao; Xu, Yong; Pei, Jian

doi:10.1109/tkde.2022.3175094

Cited by 20 publications

(11 citation statements)

References 61 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1) Methods in MBSR: MBSR needs to model the sequence and behavior types at the same time, and it also needs to consider long-term and short-term preferences, as well as local or global information, which provides opportunities for employing different technologies. In MBSR, there are some works utilizing different techniques, including MKM-SR [15], MBGNN [75], MBHT [91], KHGT [92] and TGT [93]. We describe some of them in detail below, and summarize the data MBHT [91] Transformer + GNN A sequence of (item, behavior) pairs…”

Section: E Hybrid-methods-based Learning Architecturementioning

confidence: 99%

Global and Personalized Graphs for Heterogeneous Sequential Recommendation by Learning Behavior Transitions and User Intentions

Chen

et al. 2022

Proceedings of the 16th ACM Conference on Recommender Systems

View full text Add to dashboard Cite

People usually have the explicit or implicit desire to get the information they need and are most interested in from massive information, which has led to the creation of personalized recommender systems. Recommender systems are set up to address the issue of information overload in traditional information retrieval systems such as search engines, and have been a significant area of research focused on recommending information that is of most interest to users. There is a sequential nature to the behavior of a person interacting with a system, such as examining one item of clothing before examining others. The problem of taking this sequential nature into account in delivering recommendation is known as sequential recommendation (SR). The traditional sequential recommendation problem merely takes into account a single type of behavior of the users, while in real scenarios users tend to engage in multiple types of behaviors, such as examining and adding clothes to cart before purchasing them, leading to the proposal of multibehavior sequential recommendation (MBSR). MBSR considers both sequentiality and heterogeneity of user behaviors, which can achieve state-of-the-art recommendation through suitable modeling. Hence, MBSR is a relatively new and worthy direction for in-depth research, for which some related works have been proposed. This survey aims to shed light on the MBSR problem. Firstly, we introduce MBSR in detail, including its problem definition, application scenarios and challenges faced. Secondly, we detail the classification of MBSR, including neighborhood-based methods, matrix factorization-based methods and deep learningbased methods, where we further classify the deep learningbased methods into different learning architectures based on RNN, GNN, Transformer, and generic architectures as well as architectures that integrate hybrid techniques. In each method, we present related works based on the data perspective and the modeling perspective, as well as analyze the strengths, weaknesses and features of these works. Finally, we discuss some promising future research directions to address the challenges and improve the current status of MBSR.

show abstract

Section: E Hybrid-methods-based Learning Architecturementioning

confidence: 99%

Global and Personalized Graphs for Heterogeneous Sequential Recommendation by Learning Behavior Transitions and User Intentions

Chen

et al. 2022

Proceedings of the 16th ACM Conference on Recommender Systems

View full text Add to dashboard Cite

show abstract

“…For instance, MBN (Shen, Ou, and Li 2022) uses a meta multi-behavior sequence encoder to model meta-knowledge across behavior sequences, and a recurring-item-aware predictor to predict duplicated items in the sequences. TGT (Xia et al 2022) utilizes a behavioraware transformer (Vaswani et al 2017) network to capture the short-term interest in multi-behavior sequences, and a temporal graph neural network to capture the multi-behavior dependencies. Although they utilize the multi-behavior sequences to capture the sequential patterns of user interests, they cannot achieve optimal performance because they do not fully consider the characteristics of multi-behavior sequences such as data imbalance and heterogeneity.…”

Section: Sequential Recommender Systemsmentioning

confidence: 99%

“…Lastly, some methods (Shen, Ou, and Li 2022;Xia et al 2022) use the sequential information in multi-behavior sequences, as mentioned above. We note that the sequential information is significant in MBRS, as a user's next item is determined as a result of the drifts of various interests revealed in the multi-behavior sequences.…”

Section: Multi-behavior Recommender Systemsmentioning

confidence: 99%

“…MBN (Shen, Ou, and Li 2022) is a multi-behavior SRS with an RNN-based meta multi-behavior sequence encoder to capture meta-knowledge between the sequences and a recurring-item-aware predictor. TGT (Xia et al 2022) is a multi-behavior SRS that captures short-term user interests with a behavior-aware transformer network and long-term user interests via a temporal graph neural network.…”

Section: Experimental Settingsmentioning

confidence: 99%

“…In realworld scenarios, however, users often take several kinds of behaviors such as click, add-to-cart, add-to-favorite, and purchase. Compared to the single-behavior data, the multibehaviors of users provide diverse perspectives of user interests, which conjointly imply the context of user interests and causal relationships between the user behaviors (Xia et al 2022). For learning such various information and effectively utilizing it to the next item recommendation, the sequence modeling for multi-behavior sequences needs to consider some unique characteristics of multi-behavior sequences.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Dynamic Multi-Behavior Sequence Modeling for Next Item Recommendation

Cho¹,

Hyun²,

Lim³

et al. 2023

Preprint

View full text Add to dashboard Cite

Sequential Recommender Systems (SRSs) aim to predict the next item that users will consume, by modeling the user interests within their item sequences. While most existing SRSs focus on a single type of user behavior, only a few pay attention to multi-behavior sequences, although they are very common in real-world scenarios. It is challenging to effectively capture the user interests within multi-behavior sequences, because the information about user interests is entangled throughout the sequences in complex relationships. To this end, we first address the characteristics of multi-behavior sequences that should be considered in SRSs, and then propose novel methods for Dynamic Multi-behavior Sequence modeling named DyMuS, which is a light version, and DyMuS + , which is an improved version, considering the characteristics. DyMuS first encodes each behavior sequence independently, and then combines the encoded sequences using dynamic routing, which dynamically integrates information required in the final result from among many candidates, based on correlations between the sequences. DyMuS + , furthermore, applies the dynamic routing even to encoding each behavior sequence to further capture the correlations at item-level. Moreover, we release a new, large and up-to-date dataset for multibehavior recommendation. Our experiments on DyMuS and DyMuS + show their superiority and the significance of capturing the characteristics of multi-behavior sequences.

show abstract

Temporal-Aware Multi-behavior Contrastive Recommendation

Xuan

2023

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Multi-Behavior Sequential Recommendation with Temporal Graph Transformer

Cited by 20 publications

References 61 publications

Global and Personalized Graphs for Heterogeneous Sequential Recommendation by Learning Behavior Transitions and User Intentions

Global and Personalized Graphs for Heterogeneous Sequential Recommendation by Learning Behavior Transitions and User Intentions

Dynamic Multi-Behavior Sequence Modeling for Next Item Recommendation

Temporal-Aware Multi-behavior Contrastive Recommendation

Contact Info

Product

Resources

About