STAR-Rec: Making Peace with Length Variance and Pattern Diversity in Sequential Recommendation

Hongzhi Yin; Maolin Wang; Ruocheng Guo; Sheng Zhang; Wanyu Wang; Xiangyu Zhao; Xuetao Wei; Yi Chang; Zitao Liu

REVIEW 5 major objections 6 minor 1 cited by

STAR-Rec: Making Peace with Length Variance and Pattern Diversity in Sequential Recommendation

T0 review · 5 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read STAR-Rec claims that a hybrid of preference-aware attention and state-space modeling, fused through a mixture-of-experts layer, beats nine sequential-recommendation baselines on four datasets.

desk verdict STAR-Rec is a plausible architecture paper with real evaluation-reporting holes; the empirical claim likely holds, but the exact margins are unverifiable as written. read the letter →

arxiv 2505.03484 v1 pith:LMXM3Z72 submitted 2025-05-06 cs.IR

Maolin Wang , Sheng Zhang , Ruocheng Guo , Wanyu Wang , Xuetao Wei , Zitao Liu , Hongzhi Yin , Yi Chang

show 1 more author

Xiangyu Zhao

This is my paper · ORCID

classification cs.IR

keywords sequentialrecommendationstate-spacemodelpreference-awareattentionmixture-of-expertslengthvariancepatterndiversityadaptivefusionnext-itemprediction

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

STAR-Rec sets out to show that a sequential recommender can handle very different sequence lengths and several coexisting behavior patterns in one architecture, instead of trading one strength for another. Its central move is to treat the state-space model and preference-aware attention as two sets of attention-like masks over the input sequence, then fuse them with adaptive weights before a mixture-of-experts layer specializes the prediction. The paper reports that this combined model outperforms nine recent baselines on four real-world datasets in Recall@10, MRR@10, and NDCG@10, with gains of 0.35% to 2.15%, and that the largest margins appear in sparse e-commerce settings with short sequences. A sympathetic reader would care because the proposal directly targets two known failure modes: SSMs weaken on short histories, and attention becomes costly on long ones.

What carries the argument

The load-bearing object is the unified mask-concatenation identity: an SSM's recursive state update can be expanded into a tensor $\mathbf{M}$ whose slices $\mathbf{M}_{d,:,:}$ act as attention-like masks over the sequence, so the SSM output becomes a concatenation of $D$ 'long-range heads,' one per feature dimension. Preference-aware multi-head attention supplies $H$ additional 'preference heads' built from query-key attention and a per-head gated cosine-similarity matrix. The fusion in equation (3) stacks these two head families and applies a block-diagonal learned weighting $\operatorname{diag}(\gamma_1 I; \gamma_2 W_O)$, so the whole model is one attention-like operator with two families of heads. The mixture-of-experts layer with a diversity-regularized gating network then selects among specialized feed-forward experts for the final prediction.

What would settle it

Re-run STAR-Rec and the nine baselines under a strictly controlled evaluation: fix one negative sampler, such as 100 random negatives per positive, and also run a full-ranking variant, using the same candidate set for every model and reporting results separately for users with short, medium, and long histories. If the 0.35% to 2.15% margins shrink, vanish, or reverse under either sampling scheme, the claimed advantage is a property of the protocol rather than of the architecture.

Watch

Extended reading notes

Core claim

The discovery the paper argues for is architectural: the selective state-space model and preference-aware multi-head attention are complementary views of the same sequence-mixing operation. In tensor form, an SSM acts as a stack of attention-like masks, one per feature dimension, that compress history through state-transition matrices; preference-aware attention builds masks from query-key interactions plus a thresholded cosine-similarity term that encodes static item relations. Equation (3) concatenates these two mask families into a single fused operator, with learned weights controlling the blend, and a gated mixture-of-experts layer then routes the fused representation to specialized predictors. Empirically, STAR-Rec reports consistent wins over nine baselines on four public datasets and shows its largest advantages when input sequences are truncated to length 5, 10, or 20, which the paper interprets as evidence that the attention path supplies the preference signal that pure SSMs miss on short histories.

Load-bearing premise

The load-bearing premise is that the reported gains come from the architecture and not from the evaluation setup: a single leave-one-out protocol with a binary cross-entropy loss whose negative-sampling and candidate-item choices are not described, which could in principle favor one model over another.

Editorial extensions

If this is right

On the four datasets tested, STAR-Rec improves Recall@10, MRR@10, and NDCG@10 over the best baseline in every configuration, which puts a hybrid SSM-attention architecture in contention for production sequential recommendation.
For short histories, the model's advantage is largest: with maximum sequence length 5, STAR-Rec reaches Recall@10 of 0.4470 versus 0.4255 for SIGMA, so cold-start and new-user settings benefit most.
A single-layer STAR-Rec performs best in the parameter study, meaning the reported accuracy does not require deep stacks and can be obtained at lower computational cost.
The mixture-of-experts and preference-aware attention are each necessary: removing them drops NDCG@10 from 0.2960 to about 0.2783 and 0.2795 respectively on the Beauty dataset, so the gains are tied to the full design.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Because both paths are expressed as attention-like masks, the same concatenation trick could be applied to other hybrid sequence models outside recommendation, such as time-series or session-based prediction, where long-range state and pairwise similarity both matter.
The paper's strongest evidence for length variance comes from truncating a dataset to maximum lengths 5, 10, and 20, but it does not state which dataset feeds Table 3; a natural follow-up is to stratify users by true sequence length and report per-group gains.
The 0.35% to 2.15% margins could partly reflect the cosine-similarity thresholding or the MoE diversity regularizer rather than the SSM-attention unification; ablating the threshold and the diversity weight separately would isolate those effects.
Since negative sampling is not described, a fair external check is to rerun all baselines with identical candidate-item pools; if the ordering persists across samplers, the architecture claim is protocol-independent.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The paper is a straightforward architecture contribution: it takes a Mamba-style state-space model, adds a preference-aware attention with per-head cosine-similarity gating and thresholding, and routes the fused representation through a small MoE. That combination is genuinely new compared with the baselines cited, and the ablations show each piece earns its keep, with PMA the most important. The fifty-seed runs are good practice, and the efficiency numbers look honest.

The main thing I'd tell you is to read the evaluation section twice. The paper never states the negative-sampling protocol or the candidate set for ranking. RecBole's default leave-one-out often uses a small set of random negatives; the high Recall@10 on Amazon (0.456 on Beauty with 12k items) strongly suggests that's what happened here. That's fine as long as every baseline is evaluated under the same protocol, and the paper says they used RecBole, so it probably was. But 'probably' isn't good enough for a paper that makes a 0.35–2.15% claim. The missing standard deviations make the significance asterisks unverifiable, Table 3 doesn't say which dataset it's on, and tau_h and lambda are never given values. Those are fixable in a revision, but they matter.

The theoretical framing is a stretch. Showing that SSM can be written as attention-like masks and then concatenated with attention heads is a structural observation, not a demonstrated unification. The paper would lose nothing by dropping the 'theoretically' language.

If the protocol details check out, this is a solid incremental contribution for the sequential-recommendation crowd. It does not open a new capability, and the gains are small, but the ablations and length-robustness experiment are useful. I'd send it to review and ask the authors to supply the missing evaluation details; with that, the empirical claim is likely to hold. Without that, the specific margins should be treated as provisional.

Referee Report

5 major / 6 minor

Summary. The paper proposes STAR-Rec, a sequential recommendation architecture that combines preference-aware multi-head attention (PMA), a state-space model (SSM/Mamba-style) path, and a mixture-of-experts (MoE) prediction layer. The authors claim that SSMs and attention can be naturally unified in recommendation scenarios, and they report consistent state-of-the-art performance on ML-1M, Amazon Beauty, Amazon Baby, and Amazon Video Games, with gains of 0.35%–2.15% over nine baselines under a RecBole leave-one-out protocol with fifty random seeds. Additional experiments examine limited historical information, ablations, hyperparameters, and efficiency.

Significance. If the empirical claim holds, STAR-Rec is a useful engineering contribution to sequential recommendation: the code is released, the architecture combines components with complementary strengths, the ablation study isolates the contribution of each module, and the efficiency comparison is informative. However, the paper's significance is currently undercut by reporting gaps. The central empirical assertion depends on a negative-sampling and candidate-set protocol that is never described, and the claimed statistical significance is not auditable because no standard deviations or confidence intervals are reported. In addition, the theoretical unification claim is not established by the presented derivation, and the key fusion equation contains a dimensionality error. These issues are fixable, but they are load-bearing for the paper's main contributions.

major comments (5)

[§4.2–4.3, Table 2] The evaluation protocol is not fully specified. The paper uses RecBole leave-one-out splitting and reports Recall@10, MRR@10, and NDCG@10, but it never states how negative items are sampled for training or how the candidate set for evaluation is constructed. This is consequential because the reported gains are small (0.35% on ML-1M Recall@10) and because the absolute metric values on the Amazon datasets are consistent with ranking against a sampled candidate set rather than the full item corpus. If different baselines were evaluated against different negative samples, or if the training negative sampler differed across models, the reported advantages could reflect protocol choices rather than architectural superiority. Please specify the negative-sampling distribution, the number of negatives, the evaluation candidate-set size and construction, the random seed, and a confirmation that every model is scored on the same candidate items.
[§4.2, Table 2] The paper states that each experiment was run with fifty random seeds and marks the best results with an asterisk for p < 0.05, but no standard deviations, confidence intervals, or p-values are reported anywhere. Without dispersion measures, the reader cannot verify whether a 0.35% Recall@10 gain on ML-1M is statistically meaningful or whether the asterisks survive a multiple-comparison correction. Please report mean ± standard deviation (or confidence intervals) for all models and all metrics, and describe the significance test used.
[§3.1.4, Eq. (3)] The central fusion equation is dimensionally inconsistent as written. The concatenated vector in Eq. (3) has column dimension D + H·D_k (D SSM slices plus H attention heads of dimension D_k), while the block-diagonal factor diag(γ1 I; γ2 W_O) with W_O ∈ R^{(H·D_k)×D} has dimensions (D + H·D_k) × 2D if interpreted as a true block-diagonal matrix. The product therefore cannot produce the D-dimensional output 𝒀_final stated in the equation. The intended computation appears to be a vertical stacking of γ1 I_D and γ2 W_O followed by a single matrix multiplication, not a block-diagonal product. Please rewrite Eq. (3) with dimensionally consistent notation.
[§3.1.4, Eqs. (1)–(2)] The claim that the paper 'theoretically demonstrates' that SSM and attention can be naturally unified is not supported by the derivation. Eq. (2) rewrites the SSM recurrence as a linear map with a tensor M, and Eq. (1) writes attention as another linear map; both are linear sequence transformations, but the derivation does not show that M satisfies the structural properties of an attention matrix (e.g., nonnegativity, row normalization, or a query–key–value factorization). Absent such conditions, the 'unification' reduces to the observation that both mechanisms can be expressed as matrix multiplications. Please either provide a formal equivalence theorem with the required structural conditions or revise the contribution statement to describe the fusion as a design choice rather than a theoretical result.
[§4.5, Table 3] The dataset used for the maximum-sequence-length experiment is never identified. Table 3 reports results for maximum lengths 5, 10, and 20 but the caption and Section 4.5 do not state which of the four datasets is used. This makes the RQ2 robustness claim impossible to interpret and prevents comparison with Table 2. The text also refers to '100 in main experiments,' while the main setup uses maximum lengths of 200 for ML-1M and 50 for the Amazon datasets; please clarify the setting and the dataset.

minor comments (6)

[§2.2] There is a typo in the SSM preliminaries: 'SSMs can described as follows' should read 'SSMs can be described as follows,' and the sentence beginning 'T The State-Space Model' contains a stray 'T'.
[§3.1.3–3.1.4] The subsection numbering and ordering are confusing: §3.1.3 is titled 'Adaptive Fusion of SSM and PMA' but contains no content, and the SSM path is presented as §3.1.4 after the fusion heading. Please reorder the subsections so that both paths are described before the fusion.
[Figure 4] The ablation figure shows point estimates only, even though the paper claims fifty-seed runs. Adding error bars or standard deviations would make the ablation claims consistent with the significance claims in Table 2.
[§4.8, Table 5] The text says the efficiency analysis includes inference time, training time, and GPU memory usage, but Table 5 contains only inference and training time columns. Either add the memory measurements or remove the mention of GPU memory.
[Throughout] The model name is inconsistently capitalized as 'STAR-Rec' and 'STAR-REC' (for example, in RQ1 and in the table headers). Please standardize the name.
[§3.1.2] The phrase 'Finall,y the user-item interaction' in the item embedding paragraph contains a typo: 'Finall,y' should be 'Finally'.

Circularity Check

0 steps flagged · score 2.0 of 10

No material circularity: STAR-Rec's gains are measured on held-out next interactions and its SSM/attention unification is a constructive re-expression, not a self-justifying derivation; minor self-citations appear only as baselines or related-work context.

full rationale

The paper's central empirical claim is that STAR-Rec outperforms nine baselines in Table 2. That claim is supported by a leave-one-out protocol on held-out interactions: 'for each user's interaction sequence, we select the penultimate interaction for validation' and 'The last interaction is used for testing', so no prediction target is used to fit the model. The theoretical claim that 'SSM and attention mechanisms can be naturally unified' is implemented by rewriting the SSM output as attention-like masks through the tensor M in Eq. (2) and then concatenating those heads with PMA heads in Eq. (3). This is a mathematical/architectural construction rather than a derivation whose conclusion was assumed; the learned weights gamma1 and gamma2 and the binary cross-entropy objective do not encode the reported improvements. The paper's self-citations (e.g., SIGMA, SMLP4Rec, LinRec) are used as experimental baselines or related-work references, not as load-bearing justification for STAR-Rec's components or results, so they do not force the empirical outcome. The score of 2 reflects only the presence of minor, non-load-bearing self-citations. Separate reproducibility concerns - unspecified negative-sampling/candidate-set procedure for the BCE metric, absence of standard deviations despite the claimed p<0.05 markers, and the unnamed dataset in Table 3 - are protocol and reporting risks, not circularity.

Assumptions & free parameters 10 free parameters · 5 assumptions · 1 invented entities

The central architecture depends on many hand-set hyperparameters, including hidden size, number of heads, number of experts, Mamba state dimension, thresholds, and loss weight, several of which are unreported. The claimed theoretical unification of SSM and attention is an assumed reformulation used to justify fusion, not a proven equivalence that yields the architecture uniquely.

free parameters (10)

hidden dimension d = 128 (ML-1M), 64 (Amazon)
Chosen per dataset and reported in Section 4.2; sensitivity analysis is given in Figure 5.
number of PMA attention heads H = 4
Set in Section 4.2; no sensitivity analysis is reported.
SSM kernel size = 4
Set in Section 4.2; not varied in the paper.
number of MoE experts = 8
Set in Section 4.2; no ablation on expert count is reported.
Mamba state dimension = 32
Set in Section 4.2.
Mamba expansion factor = 2
Set in Section 4.2.
max sequence length = 200 (ML-1M), 50 (Amazon)
Set in Section 4.2; RQ2 varies it to 5, 10, and 20.
dropout rate = 0.2 (ML-1M), 0.5 (Amazon)
Set in Section 4.2.
diversity loss weight lambda = not reported
Appears in L = L_pred + lambda L_div in Section 3.2; its value is not given.
preference-aware threshold tau_h = not reported
Used in the thresholding equation in Section 3.1.2; whether it is fixed, learned, or tuned per head is not stated.

assumptions (5)

standard math SSM discretization formulas, Pbar = exp(Delta P) and Qbar = (Delta P)^-1(exp(Delta P)-I)Delta Q, are valid and stable for the Mamba backbone.
Taken from the SSM and Mamba literature in Section 2.2; this is standard background.
domain assumption Leave-one-out splitting with the penultimate and last items as validation and test is a fair proxy for next-item prediction quality.
Section 4.1; this protocol is common in the cited literature, but it is an evaluation assumption that can favor certain architectures.
domain assumption Binary cross-entropy on implicit feedback with sampled negatives is a valid training objective.
Section 3.2 defines the loss but does not specify the negative sampler, so the comparison rests on an unstated protocol shared by all models.
ad hoc to paper The tensor factorization in Eq. (2) fully characterizes the SSM path, allowing SSM to be treated as an attention mechanism with D heads.
Section 3.1.4; this reformulation is used to justify the concatenation in Eq. (3). It is a design argument, not a derived theorem.
domain assumption The 0.01 floor in the thresholded similarity matrix preserves differentiability and numerical stability.
Section 3.1.2; no mathematical or empirical justification is given for the specific constant.

invented entities (1)

Preference-aware Multi-head Attention with head-specific gating, cosine-similarity matrix, and per-head threshold
purpose: Capture static item similarities and user preference patterns that the SSM path is claimed to miss.
The only evidence for this component is the ablation in Figure 4 on the Beauty dataset and the final performance tables; no external benchmark or formal analysis isolates the mechanism.

how reviews work

0 comments

Cite this review

Pith. "Pith review of STAR-Rec: Making Peace with Length Variance and Pattern Diversity in Sequential Recommendation." pith.science (2026). https://pith.science/paper/LMXM3Z72

@misc{pith2026250503484,
  author       = {Pith},
  title        = {Pith review of: STAR-Rec: Making Peace with Length Variance and Pattern Diversity in Sequential Recommendation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LMXM3Z72}},
  note         = {Machine review of arXiv:2505.03484}
}

read the original abstract

Recent deep sequential recommendation models often struggle to effectively model key characteristics of user behaviors, particularly in handling sequence length variations and capturing diverse interaction patterns. We propose STAR-Rec, a novel architecture that synergistically combines preference-aware attention and state-space modeling through a sequence-level mixture-of-experts framework. STAR-Rec addresses these challenges by: (1) employing preference-aware attention to capture both inherently similar item relationships and diverse preferences, (2) utilizing state-space modeling to efficiently process variable-length sequences with linear complexity, and (3) incorporating a mixture-of-experts component that adaptively routes different behavioral patterns to specialized experts, handling both focused category-specific browsing and diverse category exploration patterns. We theoretically demonstrate how the state space model and attention mechanisms can be naturally unified in recommendation scenarios, where SSM captures temporal dynamics through state compression while attention models both similar and diverse item relationships. Extensive experiments on four real-world datasets demonstrate that STAR-Rec consistently outperforms state-of-the-art sequential recommendation methods, particularly in scenarios involving diverse user behaviors and varying sequence lengths.

Figures

Figures reproduced from arXiv: 2505.03484 by the authors.

**Figure 1.** Overview of STAR-Rec architecture. The model consists of (1) an item embedding layer for initial representations, (2) a [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗

**Figure 2.** Overview of the adaptive fusion mechanism inte [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Illustration of the unified computational pattern [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Ablation study of different components in our [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

**Figure 5.** Figure 5: Impact of hidden size 𝑑 on model performance. The experimental results provide strong evidence for the effectiveness of our design choices. First, removing the MoE prediction layer leads to a noticeable performance drop (e.g., NDCG@10 decreases from 0.2960 to 0.2783)…

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

Block-Biased Mamba for Long-Range Sequence Processing
cs.LG 2025-05 conditional novelty 6.0 of 10

B2S6, a block-biased variant of Mamba's S6 unit, restores universal approximation, softens input-dependent forgetting, and beats S4/S4D on Long-Range Arena while keeping Mamba-like language perplexity.

Reference graph

Works this paper leans on

65 extracted references · 30 canonical work pages · cited by 1 Pith paper

[1]

Tesfaye Fenta Boka, Zhendong Niu, and Rama Bastola Neupane. 2024. A sur- vey of sequential recommendation systems: Techniques, evaluation, and future directions. Information Systems (2024)

work page 2024
[2]

Yang Cao and Wei Zhang. 2024. Mamba4KT: An Efficient and Effective Mamba- based Knowledge Tracing Model. arXiv preprint arXiv:2405.16542 (2024)

arXiv 2024
[3]

Jianxin Chang, Chenbin Zhang, Yiqun Hui, Dewei Leng, Yanan Niu, Yang Song, and Kun Gai. 2023. Pepnet: Parameter and embedding personalized network for infusing with personalized prior information. In Proc. of KDD

work page 2023
[4]

Xiaoqing Chen, Zhitao Li, Weike Pan, and Zhong Ming. 2023. A Survey on Multi-Behavior Sequential Recommendation. arXiv preprint arXiv:2308.15701 (2023)

arXiv 2023
[5]

Junyoung Chung, Caglar Gulcehre, KyungHyun Cho, and Yoshua Bengio. 2014. Empirical evaluation of gated recurrent neural networks on sequence modeling. arXiv preprint arXiv:1412.3555 (2014)

arXiv 2014
[6]

Ziqiang Cui, Haolun Wu, Bowei He, Ji Cheng, and Chen Ma. 2024. Context Matters: Enhancing Sequential Recommendation with Context-aware Diffusion- based Contrastive Learning. In Proceedings of the 33rd ACM International Confer- ence on Information and Knowledge Management . 404–414

2024
[7]

Tri Dao and Albert Gu. 2024. Transformers are SSMs: Generalized models and efficient algorithms through structured state space duality. arXiv preprint arXiv:2405.21060 (2024)

arXiv 2024
[8]

Rahul Dey and Fathi M Salem. 2017. Gate-variants of gated recurrent unit (GRU) neural networks. In 2017 IEEE 60th international midwest symposium on circuits and systems (MWSCAS)

work page 2017

Show all 65 references

[9]

Xinyu Du, Huanhuan Yuan, Pengpeng Zhao, Jianfeng Qu, Fuzhen Zhuang, Guan- feng Liu, Yanchi Liu, and Victor S Sheng. 2023. Frequency enhanced hybrid attention network for sequential recommendation. In Proc. of SIGIR

2023
[10]

Xinyan Fan, Zheng Liu, Jianxun Lian, Wayne Xin Zhao, Xing Xie, and Ji-Rong Wen. 2021. Lighter and better: low-rank decomposed self-attention networks for next-item recommendation. In Proc. of SIGIR

2021
[11]

Jingtong Gao, Xiangyu Zhao, Muyang Li, Minghao Zhao, Runze Wu, Ruocheng Guo, Yiding Liu, and Dawei Yin. 2024. SMLP4Rec: an Efficient all-MLP architec- ture for sequential recommendations. ACM Transactions on Information Systems 42, 3 (2024), 1–23

2024
[12]

Jingtong Gao, Xiangyu Zhao, Muyang Li, Minghao Zhao, Runze Wu, Ruocheng Guo, Yiding Liu, and Dawei Yin. 2024. SMLP4Rec: An Efficient all-MLP Architec- ture for Sequential Recommendations. ACM TOIS (2024)

2024
[13]

Xavier Glorot and Yoshua Bengio. 2010. Understanding the difficulty of training deep feedforward neural networks. In AISTATS

2010
[14]

Albert Gu and Tri Dao. 2023. Mamba: Linear-time sequence modeling with selective state spaces. arXiv preprint arXiv:2312.00752 (2023)

2023 arXiv
[15]

James D Hamilton. 1994. State-space models. Handbook of econometrics (1994)

1994
[16]

Balázs Hidasi, Alexandros Karatzoglou, Linas Baltrunas, and Domonkos Tikk
[17]

Xilin Jiang, Cong Han, and Nima Mesgarani. 2024. Dual-path mamba: Short and long-term bidirectional selective structured state space models for speech separation. arXiv preprint arXiv:2403.18257 (2024)

2024 arXiv
[18]

Wang-Cheng Kang and Julian McAuley. 2018. Self-attentive sequential recom- mendation. In ICDM

2018
[19]

Feyza Duman Keles, Pruthuvi Mahesakya Wijewardena, and Chinmay Hegde
[20]

Diederik P Kingma and Jimmy Ba. 2014. Adam: A method for stochastic opti- mization. arXiv preprint arXiv:1412.6980 (2014)

2014 arXiv
[21]

Yehuda Koren, Robert Bell, and Chris Volinsky. 2009. Matrix factorization tech- niques for recommender systems. Computer (2009)

2009
[22]

Chengxi Li, Yejing Wang, Qidong Liu, Xiangyu Zhao, Wanyu Wang, Yiqi Wang, Lixin Zou, Wenqi Fan, and Qing Li. 2023. STRec: Sparse Transformer for Sequen- tial Recommendations. InProceedings of the 17th ACM Conference on Recommender Systems

2023
[23]

Jing Li, Pengjie Ren, Zhumin Chen, Zhaochun Ren, Tao Lian, and Jun Ma. 2017. Neural attentive session-based recommendation. In Proceedings of the 2017 ACM on Conference on Information and Knowledge Management

2017
[24]

Muyang Li, Xiangyu Zhao, Chuan Lyu, Minghao Zhao, Runze Wu, and Ruocheng Guo. 2022. MLP4Rec: A pure MLP architecture for sequential recommendations. arXiv preprint arXiv:2204.11510 (2022)

2022 arXiv
[25]

Jiahao Liang, Xiangyu Zhao, Muyang Li, Zijian Zhang, Wanyu Wang, Haochen Liu, and Zitao Liu. 2023. Mmmlp: Multi-modal multilayer perceptron for sequen- tial recommendations. InProceedings of the ACM Web Conference 2023. 1109–1117

2023
[26]

Chengkai Liu, Jianghao Lin, Jianling Wang, Hanzhou Liu, and James Caverlee
[27]

Langming Liu, Liu Cai, Chi Zhang, Xiangyu Zhao, Jingtong Gao, Wanyu Wang, Yifu Lv, Wenqi Fan, Yiqi Wang, Ming He, et al. 2023. Linrec: Linear attention mechanism for long-term sequential recommender systems. In Proc. of SIGIR

2023
[28]

Sijia Liu, Jiahao Liu, Hansu Gu, Dongsheng Li, Tun Lu, Peng Zhang, and Ning Gu. 2023. Autoseqrec: Autoencoder for efficient sequential recommendation. In Proceedings of the 32nd ACM CIKM

2023
[29]

Ziwei Liu, Qidong Liu, Yejing Wang, Wanyu Wang, Pengyue Jia, Maolin Wang, Zitao Liu, Yi Chang, and Xiangyu Zhao. 2024. Bidirectional gated mamba for sequential recommendation. arXiv preprint arXiv:2408.11451 (2024)

2024 arXiv
[30]

Ziru Liu, Shuchang Liu, Zijian Zhang, Qingpeng Cai, Xiangyu Zhao, Kesen Zhao, Lantao Hu, Peng Jiang, and Kun Gai. 2024. Sequential recommendation for optimizing both immediate feedback and long-term retention. In Proceedings of the 47th International ACM SIGIR Conference on Re...

2024
[31]

Chao Long, Huanhuan Yuan, Junhua Fang, Xuefeng Xian, Guanfeng Liu, Victor S Sheng, and Pengpeng Zhao. 2024. Learning Global and Multi-granularity Local Representation with MLP for Sequential Recommendation. ACM Transactions on Knowledge Discovery from Data (2024)

2024
[32]

Haohao Qu, Liangbo Ning, Rui An, Wenqi Fan, Tyler Derr, Hui Liu, Xin Xu, and Qing Li. 2024. A survey of mamba. arXiv preprint arXiv:2408.01129 (2024)

2024 arXiv
[33]

Massimo Quadrana, Alexandros Karatzoglou, Balázs Hidasi, and Paolo Cremonesi
[34]

Guizhu Shen, Qingping Tan, Haoyu Zhang, Ping Zeng, and Jianjun Xu. 2018. Deep learning with gated recurrent unit networks for financial sequence predictions. Procedia computer science (2018)

2018
[35]

Fei Sun, Jun Liu, Jian Wu, Changhua Pei, Xiao Lin, Wenwu Ou, and Peng Jiang
[36]

Qiaoyu Tan, Jianwei Zhang, Ninghao Liu, Xiao Huang, Hongxia Yang, Jingren Zhou, and Xia Hu. 2021. Dynamic memory based attention network for sequential recommendation. In Proc. of AAAI

2021
[37]

Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. Proc. of NeurIPS (2017)

2017
[38]

Hanbing Wang, Xiaorui Liu, Wenqi Fan, Xiangyu Zhao, Venkataramana Kini, Devendra Yadav, Fei Wang, Zhen Wen, Jiliang Tang, and Hui Liu. 2024. Rethinking large language model architectures for sequential recommendations. arXiv preprint arXiv:2402.09543 (2024)

2024 arXiv
[39]

Maolin Wang, Yao Zhao, Jiajia Liu, Jingdong Chen, Chenyi Zhuang, Jinjie Gu, Ruocheng Guo, and Xiangyu Zhao. 2024. Large multimodal model compression via iterative efficient pruning and distillation. In Companion Proceedings of the ACM Web Conference 2024. 235–244

2024
[40]

Sinong Wang, Belinda Z Li, Madian Khabsa, Han Fang, and Hao Ma. 2020. Lin- former: Self-attention with linear complexity. arXiv preprint arXiv:2006.04768 (2020)

2020 arXiv
[41]

Yuda Wang, Xuxin He, and Shengxin Zhu. 2024. EchoMamba4Rec: Harmonizing Bidirectional State Space Models with Spectral Filtering for Advanced Sequential Recommendation. arXiv preprint arXiv:2406.02638 (2024)

2024 arXiv
[42]

Yuhao Wang, Ha Tsz Lam, Yi Wong, Ziru Liu, Xiangyu Zhao, Yichao Wang, Bo Chen, Huifeng Guo, and Ruiming Tang. 2023. Multi-task deep recommender systems: A survey. arXiv preprint arXiv:2302.03525 (2023)

2023 arXiv
[43]

Musen Wen, Deepak Kumar Vasthimal, Alan Lu, Tian Wang, and Aimin Guo. 2019. Building large-scale deep learning system for entity recognition in e-commerce search. In Proceedings of the 6th IEEE/ACM International Conference on Big Data Computing, Applications and Technologies

2019
[44]

Liwei Wu, Shuqing Li, Cho-Jui Hsieh, and James Sharpnack. 2020. SSE-PT: Sequential recommendation via personalized transformer. In RecSys

2020
[45]

Lanling Xu, Zhen Tian, Gaowei Zhang, Junjie Zhang, Lei Wang, Bowen Zheng, Yifan Li, Jiakai Tang, Zeyu Zhang, Yupeng Hou, Xingyu Pan, Wayne Xin Zhao, Xu Chen, and Ji-Rong Wen. 2023. Towards a More User-Friendly and Easy-to-Use Benchmark Library for Recommender Systems. In Proc....

2023
[47]

Jiyuan Yang, Yuanzi Li, Jingyu Zhao, Hanbing Wang, Muyang Ma, Jun Ma, Zhaochun Ren, Mengqi Zhang, Xin Xin, Zhumin Chen, et al. 2024. Uncovering Se- lective State Space Model’s Capabilities in Lifelong Sequential Recommendation. arXiv preprint arXiv:2403.16371 (2024)

2024 arXiv
[48]

Annan Yu, Michael W Mahoney, and N Benjamin Erichson. 2024. There is HOPE to Avoid HiPPOs for Long-memory State Space Models. arXiv preprint arXiv:2405.13975 (2024)

2024 arXiv
[49]

Enming Yuan, Wei Guo, Zhicheng He, Huifeng Guo, Chengkai Liu, and Ruiming Tang. 2022. Multi-behavior sequential transformer recommender. In Proc. of SIGIR

2022
[50]

Chi Zhang, Yantong Du, Xiangyu Zhao, Qilong Han, Rui Chen, and Li Li. 2022. Hierarchical item inconsistency signal learning for sequence denoising in se- quential recommendation. In Proceedings of the 31st ACM international conference on information & knowledge management . 25...

2022
[51]

Sheng Zhang, Maolin Wang, and Xiangyu Zhao. 2024. GLINT-RU: Gated Light- weight Intelligent Recurrent Units for Sequential Recommender Systems. arXiv preprint arXiv:2406.10244 (2024)

2024 arXiv
[52]

Sheng Zhang, Maolin Wang, Xiangyu Zhao, Ruocheng Guo, Yao Zhao, Chenyi Zhuang, Jinjie Gu, Zijian Zhang, and Hongzhi Yin. 2024. DNS-Rec: Data-aware Neural Architecture Search for Recommender Systems. In Proceedings of the 18th ACM Conference on Recommender Systems . 591–600

2024
[53]

Shuai Zhang, Lina Yao, Aixin Sun, and Yi Tay. 2019. Deep learning based recom- mender system: A survey and new perspectives. CSUR (2019)

2019
[54]

Kesen Zhao, Lixin Zou, Xiangyu Zhao, Maolin Wang, and Dawei Yin. 2023. User retention-oriented recommendation with decision transformer. In Proceedings of the ACM Web Conference 2023

2023
[55]

Wayne Xin Zhao, Shanlei Mu, Yupeng Hou, Zihan Lin, Yushuo Chen, Xingyu Pan, Kaiyuan Li, Yujie Lu, Hui Wang, Changxin Tian, Yingqian Min, Zhichao Feng, Xinyan Fan, Xu Chen, Pengfei Wang, Wendi Ji, Yaliang Li, Xiaoling Wang, and Ji-Rong Wen. 2021. RecBole: Towards a Unified, Com...

2021
[57]

Xiangyu Zhao, Maolin Wang, Xinjian Zhao, Jiansheng Li, Shucheng Zhou, Dawei Yin, Qing Li, Jiliang Tang, and Ruocheng Guo. 2023. Embedding in Recommender Systems: A Survey. arXiv preprint arXiv:2310.18608 (2023)

2023
[58]

Xiangyu Zhao, Long Xia, Liang Zhang, Zhuoye Ding, Dawei Yin, and Jiliang Tang. 2018. Deep reinforcement learning for page-wise recommendations. In Proceedings of the 12th ACM conference on recommender systems . 95–103

2018
[59]

Xiangyu Zhao, Liang Zhang, Zhuoye Ding, Long Xia, Jiliang Tang, and Dawei Yin
[60]

Kun Zhou, Hui Yu, Wayne Xin Zhao, and Ji-Rong Wen. 2022. Filter-enhanced MLP is all you need for sequential recommendation. In Proceedings of the ACM web conference 2022

2022
[61]

Lianghui Zhu, Bencheng Liao, Qian Zhang, Xinlong Wang, Wenyu Liu, and Xinggang Wang. 2024. Vision mamba: Efficient visual representation learning with bidirectional state space model. arXiv preprint arXiv:2401.09417 (2024)

2024 arXiv
[2015]

arXiv preprint arXiv:1511.06939 (2015)

Session-based recommendations with recurrent neural networks. arXiv preprint arXiv:1511.06939 (2015)

2015 arXiv
[2017]

In proceedings of the Eleventh ACM Conference on Recommender Systems

Personalizing session-based recommendations with hierarchical recurrent neural networks. In proceedings of the Eleventh ACM Conference on Recommender Systems
[2018]

In Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining

Recommendations with negative feedback via pairwise deep reinforcement learning. In Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining . 1040–1048
[2019]

BERT4Rec: Sequential recommendation with bidirectional encoder repre- sentations from transformer. In Proc. of CIKM
[2023]

On the computational complexity of self-attention. In Proc. of ALT
[2024]

arXiv preprint arXiv:2403.03900 (2024)

Mamba4Rec: Towards Efficient Sequential Recommendation with Selective State Space Models. arXiv preprint arXiv:2403.03900 (2024)

2024 arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.

[1] [1]

Tesfaye Fenta Boka, Zhendong Niu, and Rama Bastola Neupane. 2024. A sur- vey of sequential recommendation systems: Techniques, evaluation, and future directions. Information Systems (2024)

work page 2024

[2] [2]

Yang Cao and Wei Zhang. 2024. Mamba4KT: An Efficient and Effective Mamba- based Knowledge Tracing Model. arXiv preprint arXiv:2405.16542 (2024)

arXiv 2024

[3] [3]

Jianxin Chang, Chenbin Zhang, Yiqun Hui, Dewei Leng, Yanan Niu, Yang Song, and Kun Gai. 2023. Pepnet: Parameter and embedding personalized network for infusing with personalized prior information. In Proc. of KDD

work page 2023

[4] [4]

Xiaoqing Chen, Zhitao Li, Weike Pan, and Zhong Ming. 2023. A Survey on Multi-Behavior Sequential Recommendation. arXiv preprint arXiv:2308.15701 (2023)

arXiv 2023

[5] [5]

Junyoung Chung, Caglar Gulcehre, KyungHyun Cho, and Yoshua Bengio. 2014. Empirical evaluation of gated recurrent neural networks on sequence modeling. arXiv preprint arXiv:1412.3555 (2014)

arXiv 2014

[6] [6]

Ziqiang Cui, Haolun Wu, Bowei He, Ji Cheng, and Chen Ma. 2024. Context Matters: Enhancing Sequential Recommendation with Context-aware Diffusion- based Contrastive Learning. In Proceedings of the 33rd ACM International Confer- ence on Information and Knowledge Management . 404–414

2024

[7] [7]

Tri Dao and Albert Gu. 2024. Transformers are SSMs: Generalized models and efficient algorithms through structured state space duality. arXiv preprint arXiv:2405.21060 (2024)

arXiv 2024

[8] [8]

Rahul Dey and Fathi M Salem. 2017. Gate-variants of gated recurrent unit (GRU) neural networks. In 2017 IEEE 60th international midwest symposium on circuits and systems (MWSCAS)

work page 2017