M2: Mixed Models with Preferences, Popularities and Transitions for Next-Basket Recommendation

Peng, Bo; Ren, Zhiyun; Parthasarathy, Srinivasan; Ning, Xia

doi:10.48550/arxiv.2004.01646

Cited by 4 publications

(7 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It has been shown [8], [10], [17] that the temporal patterns play an important role in predicting users' preferences.…”

Section: Position-sensitive Preference Prediction (Ps)mentioning

confidence: 99%

“…Existing session-based recommendation methods model the temporal patterns primarily using GRUs-based methods [7], [23], which implicitly learn weights over items. However, as demonstrated in the literature [8], the complicated GRUsbased models may not be well learned for the notoriously sparse recommendation datasets, and it also suffers from poor interpretability [24] and limited parallelizability [8], [19].…”

Section: Position-sensitive Preference Prediction (Ps)mentioning

confidence: 99%

“…• Tmall (TM) 7 is from the IJCAI-15 competition, and it includes anonymized shopping logs on the online retail platform Tmall. Following the literature [1], [2], [3], [4], for GA, we only keep the top 30,000 most popular locations, and view the check-in records in one day as a session [4], [8]. For LF, we only keep the top 40,000 most popular artists, and view the listening events in 8 hours as a session [4].…”

Section: Datasetsmentioning

confidence: 99%

“…In addition, existing methods primarily model the temporal patterns using gated recurrent units (GRUs) [7]. However, considering the notorious sparse nature of session-based recommendation datasets, as shown in the literature [8], the complicated GRUs may not be well-learned and could degrade the performance. Due to its recurrent essence, GRUs also suffer from limited parallelizability and poor interpretability.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Prospective Preference Enhanced Mixed Attentive Model for Session-based Recommendation

Peng¹,

Tai²,

Parthasarathy³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Session-based recommendation aims to generate recommendations for the next item of users' interest based on a given session. In this manuscript, we develop prospective preference enhanced mixed attentive model (P 2 MAM) to generate session-based recommendations using two important factors: temporal patterns and estimates of users' prospective preferences. Unlike existing methods, P 2 MAM models the temporal patterns using a light-weight while effective position-sensitive attention mechanism. In P 2 MAM, we also leverage the estimate of users' prospective preferences to signify important items, and generate better recommendations. Our experimental results demonstrate that P 2 MAM models significantly outperform the state-of-the-art methods in six benchmark datasets, with an improvement as much as 19.2%. In addition, our run-time performance comparison demonstrates that during testing, P 2 MAM models are much more efficient than the best baseline method, with a significant average speedup of 47.7 folds.

show abstract

“…It has been shown [8], [10], [17] that the temporal patterns play an important role in predicting users' preferences.…”

Section: Position-sensitive Preference Prediction (Ps)mentioning

confidence: 99%

Section: Position-sensitive Preference Prediction (Ps)mentioning

confidence: 99%

Section: Datasetsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Prospective Preference Enhanced Mixed Attentive Model for Session-based Recommendation

Peng¹,

Tai²,

Parthasarathy³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Recommender system has become essential in providing personalized information filtering services in a variety of applications [21,26,31,40,41]. It learns the user and item embeddings from historical records on the user-item interactions [8,34].…”

Section: Introductionmentioning

confidence: 99%

Continuous-Time Sequential Recommendation with Temporal Graph Collaborative Transformer

Fan

Liu

Zhang

et al. 2021

Proceedings of the 30th ACM International Conference on Information &Amp; Knowledge Management

146

View full text Add to dashboard Cite

In order to model the evolution of user preference, we should learn user/item embeddings based on time-ordered item purchasing sequences, which is defined as Sequential Recommendation (SR) problem. Existing methods leverage sequential patterns to model item transitions. However, most of them ignore crucial temporal collaborative signals, which are latent in evolving user-item interactions and coexist with sequential patterns. Therefore, we propose to unify sequential patterns and temporal collaborative signals to improve the quality of recommendation, which is rather challenging. Firstly, it is hard to simultaneously encode sequential patterns and collaborative signals. Secondly, it is non-trivial to express the temporal effects of collaborative signals.Hence, we design a new framework Temporal Graph Sequential Recommender (TGSRec) upon our defined continuous-time bipartite graph. We propose a novel Temporal Collaborative Transformer (TCT) layer in TGSRec, which advances the self-attention mechanism by adopting a novel collaborative attention. TCT layer can simultaneously capture collaborative signals from both users and items, as well as considering temporal dynamics inside sequential patterns. We propagate the information learned from TCT layer over the temporal graph to unify sequential patterns and temporal collaborative signals. Empirical results on five datasets show that TGSRec significantly outperforms other baselines, in average up to 22.5% and 22.1% absolute improvements in Recall@10 and MRR, respectively. Our code is available online in https://github.com/ DyGRec/TGSRec.

show abstract

Continuous-Time Sequential Recommendation with Temporal Graph Collaborative Transformer

Fan¹,

Liu²,

Zhang³

et al. 2021

Preprint

View full text Add to dashboard Cite

M2: Mixed Models with Preferences, Popularities and Transitions for Next-Basket Recommendation

Cited by 4 publications

References 24 publications

Prospective Preference Enhanced Mixed Attentive Model for Session-based Recommendation

Prospective Preference Enhanced Mixed Attentive Model for Session-based Recommendation

Continuous-Time Sequential Recommendation with Temporal Graph Collaborative Transformer

Continuous-Time Sequential Recommendation with Temporal Graph Collaborative Transformer

Contact Info

Product

Resources

About