Pith. sign in

REVIEW 4 major objections 6 minor 43 references

PERSCEN: Learning Personalized Interaction Pattern and Scenario Preference for Multi-Scenario Matching

T0 review · 4 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 User-specific feature graphs and vector-quantized scenario preferences improve multi-scenario matching over shared-pattern baselines, the paper claims.

desk verdict PERSCEN is a clean, well-ablationed integration of EmerG-style user-specific graphs, VQ-based scenario preference, and a GLU fusion for multi-scenario matching; the gains are consistent but modest, and the central 'user-specific graph' claim is the softest spot. read the letter →

arxiv 2506.18382 v1 pith:Y2YSAVN4 submitted 2025-06-23 cs.IR cs.AIcs.LG

Haotong Du , Yaqing Wang , Fei Xiong , Lei Shao , Ming Liu , Hao Gu , Quanming Yao , Zhen Wang This is my paper · ORCID
classification cs.IRcs.AIcs.LG
keywords multi-scenariomatchingrecommendersystemsuser-specificmodelingfeaturegraphvectorquantizationscenario-awarepreferencetwo-towerretrievalgatedlinearunit
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

This paper argues that multi-scenario matching models have been missing the user: by modulating behavior with scenario context alone, prior methods learn a common interaction pattern for everyone. PERSCEN introduces user-specific modeling into the matching stage, building a per-user feature graph whose edges are generated from that user's own profile fields and distilling scenario-aware preferences from the user's behavior in each scenario through a shared vector-quantized codebook. On the KuaiRand-Pure and Alimama datasets, the model reports consistent Recall@K and Hits@K improvements over single- and multi-scenario two-tower baselines, with the largest relative gains in the sparsest scenarios. The point matters because matching must stay cheap enough for industrial retrieval, so the claim is that personalization can be added without abandoning two-tower efficiency.

What carries the argument

The load-bearing object is the user-specific feature graph. For each user $u$, a field-wise graph generator computes the $m$-th row of an adjacency matrix $A_u^{(1)}$ as $\mathrm{MLP}_m([e_{u,1},\ldots,e_{u,N_f},\mathrm{onehot}(m)])$, so edges depend on the user's own feature embeddings; higher-order structure is obtained by matrix products and refined by normalization, sparsification, and symmetrization. A message-free GNN update $h_{u,m}^{(l)} = h_{u,m}^{(l-1)} \odot \sum_{n=1}^{N_f} [\bar{A}_u^{(l-1)}]_{mn} W_g^{(l-1)} h_{u,n}^{(0)}$ propagates over this graph to yield the shared-preference representation. Scenario-aware preferences come from vector quantization: a latent $z_{u,s}$ from the scenario-specific behavior sequence is matched to the nearest codebook vector $c_j$, giving $p_{u,s} = z_{u,s} + c_j$, trained with a straight-through estimator and a VQ loss. The progressive scenario-aware GLU $g_{u,s}^{(l)} = (W_{r1}[h_u^{(l)}, g_{u,s}^{(l-1)}] + W_{r2}\hat{p}_{u,s}) \odot \sigma(W_{r3}[h_u^{(l)}, g_{u,s}^{(l-1)}] + W_{r4}\hat{p}_{u,s})$ adaptively fuses the two kinds of information. These pieces together carry the claim that personalization can be done with matrix products and gating rather than heavy sequence models.

What would settle it

One concrete check: train PERSCEN on a dataset where user IDs are randomly permuted so profiles no longer match behavior sequences, and measure Recall@K. If performance does not drop materially, the user-specific graph is not the source of the gains. A second check: compute the pairwise similarity or entropy of learned adjacency matrices across users; near-identical matrices would indicate the generator has collapsed to a shared graph.

Watch

Extended reading notes

Core claim

The paper's central claim is that incorporating user-specific modeling into multi-scenario matching improves retrieval accuracy over existing approaches that rely on scenario context alone. PERSCEN generates a user-specific adjacency matrix for each user by applying a per-field MLP to that user's feature embeddings; a lightweight GNN then propagates over this matrix to capture higher-order feature interactions that are shared across scenarios but personalized. Scenario-aware preferences are obtained by vector-quantizing the user's scenario-specific behavior sequence against a codebook shared by all scenarios, so preferences learned in data-rich scenarios can transfer to data-sparse ones. A progressive scenario-aware gated linear unit fuses the shared and scenario-specific representations. The experiments report that PERSCEN outperforms the compared baselines in scenario-wise Recall@K and Hits@K on both datasets, and the paper states that it is the first to incorporate user-specific modeling into multi-scenario matching.

Load-bearing premise

The load-bearing premise is that a network that sees only a user's own profile fields can produce a genuinely user-specific graph of feature interactions; if the learned adjacency matrices end up nearly identical across users, the personalization collapses to the shared-graph variant that the paper's own ablation shows to be weaker.

Editorial extensions

If this is right

  • If PERSCEN's claims hold, multi-scenario matching models can personalize shared preferences per user without leaving the two-tower architecture, since the added components are matrix products and gating.
  • In data-sparse scenarios, the shared codebook acts as a transfer channel: preferences learned in large scenarios can be assigned to users in small scenarios, which is where the reported gains are largest.
  • The ablation results imply each of the three components, the user-specific graph, the VQ-based scenario preference, and the progressive GLU, contributes independently, since removing any one lowers Recall@K.
  • The efficiency analysis suggests the personalization cost is bounded: training time grows relative to baselines but FLOPs remain at roughly 8 to 13 GFLOPs, within the stated industrial deployment range.
  • For practitioners, the model provides a template for adding fine-grained user modeling to matching without moving to ranking-style heavy networks.

Reading between the lines

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

  • The user-specific adjacency matrices could be reused as interpretable user profiles or as inputs to downstream ranking, since they already encode which profile fields interact for each user; the paper only visualizes them for six users.
  • The VQ codebook size acts as a granularity dial for scenario-aware preferences, and the sensitivity analysis shows an optimum at intermediate sizes; one could test whether the optimal size scales with the number of scenarios or with behavior-sequence diversity.
  • A natural extension not tested here is cold-start users with very short scenario-specific sequences, where the codebook's shared prior should help most; this could be measured by slicing results by sequence length.
  • Because the method is designed for the matching stage, the same user-specific graph and VQ components could be dropped into the item tower or into a ranking model, though the efficiency claims would need re-checking there.
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.

Referee Report

4 major / 6 minor

Summary. The paper proposes PERSCEN, a multi-scenario matching model that adds user-specific modeling to the two-tower matching paradigm. It builds a per-user feature graph via a field-wise MLP that outputs a user-specific adjacency matrix, applies a lightweight GNN to capture higher-order feature interactions, extracts scenario-aware preferences by vector-quantizing scenario-specific behavior sequences, and fuses the two via a progressive scenario-aware GLU. The method is evaluated on KuaiRand-Pure and Alimama with Recall@K and Hits@K, and the authors report consistent gains over single- and multi-scenario baselines, plus efficiency and ablation analyses. The writing is clear and the code is released, but the empirical support for the two personalized components is largely qualitative and some comparisons lack statistical testing.

Significance. If the claims hold, PERSCEN is a useful contribution to multi-scenario retrieval: it is a practical two-tower architecture that injects user-specific feature interactions and scenario-aware preference codes without heavy latency. The method is clearly specified, the ablations isolate each component, the efficiency numbers are reported, and the code availability is a strength. There is no circularity concern: the losses are standard task and VQ losses, and hyperparameters are selected on a validation split. The main risk is evidential: the central novelty rests on the user-specific adjacency generator and the shared codebook, but the paper offers only qualitative visualizations and small, sometimes non-significant ablation gaps to show that these mechanisms work as intended. The inconsistent codebook-size report also needs correction.

major comments (4)
  1. [Table 4 and Table 5] No statistical significance tests are reported, and several key gaps are small relative to the reported standard deviations. For example, on KuaiRand K3, PERSCEN's Recall@50 is 19.60±0.86 versus ADIN's 18.88±0.62, and the Recall@100 gap is 30.50±0.89 versus 29.82±1.37; both differences are within one standard deviation. In Table 5, the w/ shared graph ablation differs from PERSCEN by only 12.48±0.49 versus 12.63±0.12 on A4 R@500. Because the central claim is that PERSCEN outperforms existing methods and that user-specific modeling is responsible for the gains, please report per-seed results, paired significance tests or confidence intervals, and a discussion of effect sizes.
  2. [Section 4.2.1, Eq. (4), Figure 4, Table 5] The field-wise feature graph generator in Eq. (4) does not compel the learned adjacency matrix to be meaningfully user-specific: the per-field MLP receives user feature embeddings and could converge to a near-common graph, a graph dominated by the user-id embedding, or a graph with only small input-dependent perturbations. The only quantitative evidence is the w/ shared graph ablation, whose differences are small and not always significant (e.g., A3 R@500: 12.03±0.41 versus 12.34±0.24), while Figure 4 shows only six hand-selected users. Please provide quantitative diagnostics: the distribution of edge weights across users, the mean pairwise similarity of adjacency matrices, a variance decomposition of A_u with respect to input features, and an ablation that removes the user-id embedding from the generator input, to demonstrate that the generated graphs are genuinely user-specific.
  3. [Section 4.3, Figure 5, Section 5.2] The claim that the shared codebook transfers preferences from data-rich to data-sparse scenarios is not supported by codebook diagnostics. The paper reports no statistics on code utilization, such as the number of active codes per scenario, the overlap of code usage across scenarios, per-user assignment entropy, or a quantitative analysis linking code usage to scenario data volume. Figure 5 is a t-SNE plot of ten users and is anecdotal. This evidence is needed because the data-sparsity improvements in Table 4 (e.g., K4 and A4) are attributed in Section 5.2 to the shared codebook mechanism.
  4. [Section B.2.2 and Table 6] There is a direct inconsistency in the reported configuration: Table 6 lists the Alimama codebook size as 5, while Section B.2.2 states that the optimal codebook size for the Alimama dataset is 10 and Figure 7 shows better performance at size 10 than at size 5. Please reconcile this discrepancy and clarify which value was used for the main results and which validation metric selected it.
minor comments (6)
  1. [Section 1] The claim that PERSCEN is 'the first to incorporate user-specific modeling into multi-scenario matching' is strong; please soften it or support it with a focused comparison to recent user-adaptive multi-scenario methods, even if those are ranking-stage, since the distinction is argued rather than demonstrated.
  2. [Eq. (1)] The notation 'arg Top_K' is nonstandard and should be replaced with a standard top-K operator or defined explicitly.
  3. [Throughout] There are several typos and grammatical slips, including 'approches' (Section 2), 'the sequal' (Section 4), 'we builds' (Section 6), 'ICAN relys' (Section 5.3), 'seatch' (Appendix A.1), and 'between between' (Section 5.5.1); a careful proofread is needed.
  4. [Section 5.1.2] The statement that K is set to approximately 1% of candidate items does not match the actual values: for Alimama with 846,811 items, K=500 is about 0.06%, not 1%. Please clarify the selection criterion.
  5. [Tables 2 and 3] The efficiency results are reported with standard deviations in Table 2 but not in Table 3, and no number of runs is given for Table 3; please state the number of runs and report variance consistently.
  6. [Figure 5] For reproducibility of the t-SNE visualization, please report the perplexity, number of iterations, and whether the plot is a single run or aggregated.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the derivation chain is self-contained and the only self-citation (EmerG) is a non-load-bearing design template.

full rationale

I walked the derivation chain from Eq. (4) through Eq. (18). Each component is a learned function of inputs, and no equation reduces to its own output by construction. In Eq. (4), the user-specific adjacency row [A_u^(1)]_{m:} is produced by an MLP over user feature embeddings plus a field indicator; the claim that different users get different graphs is a property of the learned parameters that the paper tests empirically via the 'w/ shared graph' ablation in Table 5, not a definitional identity. The VQ path in Eqs. (7)-(10) learns z_{u,s} and a shared codebook jointly; the codebook-loss/commitment-loss objective is a standard regularizer, and the sparse-scenario transfer claim is an empirical benefit of sharing codes, not a restatement of the metric. Eq. (13) fuses previously computed h_u and p-hat_{u,s}; every input is an upstream learned representation, and the task loss (17) plus VQ loss (10) is optimized on held-out test metrics, with hyperparameters selected on validation. The only self-citation is EmerG [27], co-authored by two of the present authors, used in Section 4.2.1 as a design template ('Inspired by EmerG [27]... follow [27]'); it is not invoked as a uniqueness theorem or as evidence for the empirical gains, which are benchmarked against external baselines ICAN, ADIN, SASS, and M5 on public datasets. There is therefore no circular step: no fitted parameter is renamed a prediction, no ansatz is smuggled in as an external fact, and no equation equals its own input by construction. The skeptics' concerns about the qualitative feature-graph evidence and near-common learned graphs are validity/robustness concerns, not circularity.

Assumptions & free parameters 9 free parameters · 5 assumptions · 0 invented entities

The central claim rests on standard matching-stage assumptions plus two ad hoc modeling premises: the user-specific graph generator produces meaningful per-user interaction topologies, and the shared VQ codebook transfers preferences across scenarios. Hyperparameters were tuned per dataset and directly affect the reported gains.

free parameters (9)
  • number of GNN layers = 3 (KuaiRand), 1 (Alimama)
    Tuned per dataset; sensitivity shown only for Alimama (Figure 6).
  • codebook size = 10 (KuaiRand), 5 (Alimama)
    Tuned; affects the granularity of scenario-aware preferences (Figure 7).
  • beta in VQ loss = 0.25
    Tuned; trades off codebook and commitment losses (Figure 8).
  • embedding dimension = 16
    Chosen from a small range; not justified by analysis.
  • batch size = 4096
    Standard choice; no sensitivity analysis.
  • learning rate = 0.001
    Standard Adam default; no sensitivity analysis.
  • weight decay = 1e-6
    Chosen from a range; no sensitivity analysis.
  • sequence pooling strategy = mean
    Chosen from sum/max/mean/concat; no sensitivity analysis.
  • number of negative samples per positive = 10
    Chosen from 1-10; no sensitivity analysis.
assumptions (5)
  • domain assumption The two-tower architecture with inner product similarity is a valid retrieval model (Eq. 16).
    Standard in industrial matching; used as the backbone throughout.
  • domain assumption Random negative sampling with binary cross-entropy loss optimizes retrieval performance (Eq. 17).
    Standard simplification; not backed by theory in the paper.
  • ad hoc to paper A per-field MLP over a user's own feature embeddings can generate a meaningful user-specific adjacency matrix, A_u (Eq. 4).
    This is the central personalization mechanism; if the MLP yields near-identical graphs across users, the method's advantage collapses to the 'w/ shared graph' ablation, which is weaker.
  • ad hoc to paper The vector quantization codebook is shared across scenarios and transfers preferences from data-rich to data-sparse scenarios.
    The transfer benefit is a stated design goal (Section 4.3) but not formally proven; the paper argues it from the shared codebook.
  • domain assumption Scenario-specific user behavior sequences, f_seq_{u,s}, are available and informative for preference distillation.
    If users have few interactions in a scenario, the VQ encoding (Eq. 7) becomes noisy; this is partially mitigated by the codebook but still assumed.

how reviews work

0 comments
Cite this review

Pith. "Pith review of PERSCEN: Learning Personalized Interaction Pattern and Scenario Preference for Multi-Scenario Matching." pith.science (2026). https://pith.science/paper/Y2YSAVN4

@misc{pith2026250618382,
  author       = {Pith},
  title        = {Pith review of: PERSCEN: Learning Personalized Interaction Pattern and Scenario Preference for Multi-Scenario Matching},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Y2YSAVN4}},
  note         = {Machine review of arXiv:2506.18382}
}
read the original abstract

With the expansion of business scales and scopes on online platforms, multi-scenario matching has become a mainstream solution to reduce maintenance costs and alleviate data sparsity. The key to effective multi-scenario recommendation lies in capturing both user preferences shared across all scenarios and scenario-aware preferences specific to each scenario. However, existing methods often overlook user-specific modeling, limiting the generation of personalized user representations. To address this, we propose PERSCEN, an innovative approach that incorporates user-specific modeling into multi-scenario matching. PERSCEN constructs a user-specific feature graph based on user characteristics and employs a lightweight graph neural network to capture higher-order interaction patterns, enabling personalized extraction of preferences shared across scenarios. Additionally, we leverage vector quantization techniques to distil scenario-aware preferences from users' behavior sequence within individual scenarios, facilitating user-specific and scenario-aware preference modeling. To enhance efficient and flexible information transfer, we introduce a progressive scenario-aware gated linear unit that allows fine-grained, low-latency fusion. Extensive experiments demonstrate that PERSCEN outperforms existing methods. Further efficiency analysis confirms that PERSCEN effectively balances performance with computational cost, ensuring its practicality for real-world industrial systems.

Figures

Figures reproduced from arXiv: 2506.18382 by the authors.

Figure 1
Figure 1. For different users, preferences shared across all sce [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Overview of existing strategies for multi-scenario matching. These methods take user profile, scenario context, and [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Framework of PERSCEN. The framework consists of three key components: (i) User-Specific Feature Graph Modeling, [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Visualizations of user-specific adjacency matrices generated by PERSCEN. [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Visualizations of user-specific and scenario-aware [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Varying the number of GNN layers for PERSCEN [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: Varying the codebook size for PERSCEN in Ali [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: Varying the 𝛽 for PERSCEN in Alimama. increases, performance first improves and then decreases. One pos￾sible reason is that as the codebook size grows, the boundaries be￾tween diverse scenario-aware preference clusters become blurred, leading to insufficient learning …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 31 canonical work pages

  1. [1]

    Yoshua Bengio, Nicholas Léonard, and Aaron Courville. 2013. Estimating or propagating gradients through stochastic neurons for conditional computation. arXiv preprint arXiv:1308.3432 (2013)

    arXiv 2013

  2. [2]

    Jiang Bian, Jizhou Huang, Shilei Ji, Yuan Liao, Xuhong Li, Qingzhong Wang, Jingbo Zhou, Dejing Dou, Yaqing Wang, and Haoyi Xiong. 2023. Feynman: Federated Learning-based Advertising for Ecosystems-Oriented Mobile Apps Recommendation. IEEE Transactions on Services Computing 16, 5 (2023), 3361– 3372

    work page 2023

  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 Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining . 3795–3804

    work page 2023

  4. [4]

    Kyunghyun Cho, Bart Van Merrienboer, Caglar Gulcehre, Dzmitry Bahdanau, Fethi Bougares, Holger Schwenk, and Yoshua Bengio. 2014. Learning phrase representations using RNN encoder–decoder for statistical machine translation. In Conference on Empirical Methods in Natural Language Processing . 1724–1734

    work page 2014

  5. [5]

    Yann N Dauphin, Angela Fan, Michael Auli, and David Grangier. 2017. Language modeling with gated convolutional networks. In International Conference on Machine Learning. PMLR, 933–941

    work page 2017

  6. [6]

    James Davidson, Benjamin Liebald, Junning Liu, Palash Nandy, Taylor Van Vleet, Ullas Gargi, Sujoy Gupta, Yu He, Mike Lambert, Blake Livingston, et al. 2010. The YouTube video recommendation system. In ACM Conference on Recommender Systems. 293–296

    work page 2010

  7. [7]

    Kun Gai, Xiaoqiang Zhu, Han Li, Kai Liu, and Zhe Wang. 2017. Learning piece- wise linear models from large scale data for ad click prediction. arXiv preprint arXiv:1704.05194 (2017)

    arXiv 2017

  8. [8]

    Chongming Gao, Shijun Li, Yuan Zhang, Jiawei Chen, Biao Li, Wenqiang Lei, Peng Jiang, and Xiangnan He. 2022. KuaiRand: An unbiased sequential recommenda- tion dataset with randomly exposed videos. In ACM International Conference on Information & Knowledge Management . 3953–3957

    work page 2022

Show all 43 references
  1. [9]

    Jingtong Gao, Bo Chen, Menghui Zhu, Xiangyu Zhao, Xiaopeng Li, Yuhao Wang, Yichao Wang, Huifeng Guo, and Ruiming Tang. 2024. HierRec: Scenario-aware hierarchical modeling for multi-scenario recommendations. InACM International Conference on Information & Knowledge Management . 653–662

    2024

  2. [10]

    Jui-Ting Huang, Ashish Sharma, Shuying Sun, Li Xia, David Zhang, Philip Pronin, Janani Padmanabhan, Giuseppe Ottaviano, and Linjun Yang. 2020. Embedding- based retrieval in facebook search. In ACM SIGKDD International Conference on Knowledge Discovery & Data Mining . 2553–2561

    2020

  3. [11]

    Po-Sen Huang, Xiaodong He, Jianfeng Gao, Li Deng, Alex Acero, and Larry Heck. 2013. Learning deep structured semantic models for web search using clickthrough data. In ACM International Conference on Information & Knowledge Management. 2333–2338

    2013

  4. [12]

    Yuchen Jiang, Qi Li, Han Zhu, Jinbei Yu, Jin Li, Ziru Xu, Huihui Dong, and Bo Zheng. 2022. Adaptive domain interest network for multi-domain recommenda- tion. In ACM International Conference on Information & Knowledge Management . 3212–3221

    2022

  5. [13]

    Jeff Johnson, Matthijs Douze, and Hervé Jégou. 2019. Billion-scale similarity search with GPUs. IEEE Transactions on Big Data 7, 3 (2019), 535–547

    2019

  6. [14]

    Diederik P Kingma. 2014. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 (2014)

    2014 arXiv

  7. [15]

    Chao Li, Zhiyuan Liu, Mengmeng Wu, Yuchi Xu, Huan Zhao, Pipei Huang, Guoliang Kang, Qiwei Chen, Wei Li, and Dik Lun Lee. 2019. Multi-interest net- work with dynamic routing for recommendation at Tmall. In ACM International Conference on Information & Knowledge Management . 2615–2623

    2019

  8. [16]

    Xiaopeng Li, Fan Yan, Xiangyu Zhao, Yichao Wang, Bo Chen, Huifeng Guo, and Ruiming Tang. 2023. HAMUR: Hyper adapter for multi-domain recommendation. In ACM International Conference on Information & Knowledge Management . 1268– 1277

    2023

  9. [17]

    Dugang Liu, Chaohua Yang, Xing Tang, Yejing Wang, Fuyuan Lyu, Weihong Luo, Xiuqiang He, Zhong Ming, and Xiangyu Zhao. 2024. MultiFS: Automated multi- scenario feature selection in deep recommender systems. In ACM International Conference on Web Search and Data Mining . 434–442

    2024

  10. [18]

    Guangyi Liu, Quanming Yao, Yongqi Zhang, and Lei Chen. 2024. Knowledge- Enhanced Recommendation with User-Centric Subgraph Network. In 2024 IEEE 40th International Conference on Data Engineering . IEEE, 1269–1281

    2024

  11. [19]

    Qijiong Liu, Xiaoyu Dong, Jiaren Xiao, Nuo Chen, Hengchang Hu, Jieming Zhu, Chenxu Zhu, Tetsuya Sakai, and Xiao-Ming Wu. 2024. Vector quantization for recommender systems: A review and outlook. arXiv preprint arXiv:2405.03110 (2024)

    2024 arXiv

  12. [20]

    Qijiong Liu, Jiaren Xiao, Lu Fan, Jieming Zhu, and Xiao-Ming Wu. 2024. Learning category trees for ID-based recommendation: exploring the power of differen- tiable vector quantization. In The Web Conference. 3521–3532

    2024

  13. [21]

    Erxue Min, Da Luo, Kangyi Lin, Chunzhen Huang, and Yang Liu. 2023. Scenario- adaptive feature interaction for click-through rate prediction. In ACM SIGKDD International Conference on Knowledge Discovery & Data Mining . 4661–4672

    2023

  14. [22]

    Qijie Shen, Wanjie Tao, Jing Zhang, Hong Wen, Zulong Chen, and Quan Lu

  15. [23]

    Hongyan Tang, Junning Liu, Ming Zhao, and Xudong Gong. 2020. Progressive lay- ered extraction (PLE): A novel multi-task learning (MTL) model for personalized recommendations. In ACM Conference on Recommender Systems . 269–278

    2020

  16. [24]

    Yu Tian, Bofang Li, Si Chen, Xubin Li, Hongbo Deng, Jian Xu, Bo Zheng, Qian Wang, and Chenliang Li. 2023. Multi-scenario ranking with adaptive feature learning. In International ACM SIGIR Conference on Research and Development in Information Retrieval. 517–526

    2023

  17. [25]

    Laurens Van der Maaten and Geoffrey Hinton. 2008. Visualizing data using t-SNE. Journal of Machine Learning Research 9, 11 (2008), 2579–2605

    2008

  18. [26]

    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

  19. [27]

    Yaqing Wang, Hongming Piao, Daxiang Dong, Quanming Yao, and Jingbo Zhou

  20. [28]

    Yuhao Wang, Yichao Wang, Zichuan Fu, Xiangyang Li, Wanyu Wang, Yuyang Ye, Xiangyu Zhao, Huifeng Guo, and Ruiming Tang. 2024. LLM4MSR: An LLM- enhanced paradigm for multi-scenario recommendation. In ACM International Conference on Information & Knowledge Management . 2472–2481

    2024

  21. [29]

    Shiguang Wu, Yaqing Wang, Qinghe Jing, Daxiang Dong, Dejing Dou, and Quan- ming Yao. 2023. ColdNAS: Search to modulate for user cold-start recommendation. In The Web Conference. 1021–1031

    2023

  22. [30]

    Ruobing Xie, Zhijie Qiu, Jun Rao, Yi Liu, Bo Zhang, and Leyu Lin. 2020. In- ternal and contextual attention network for cold-start multi-channel matching in recommendation.. In International Joint Conference on Artificial Intelligence . 2732–2738

    2020

  23. [31]

    Jing Yan, Liu Jiang, Jianfei Cui, Zhichen Zhao, Xingyan Bin, Feng Zhang, and Zuotao Liu. 2024. Trinity: Syncretizing Multi-/Long-Tail/Long-Term Interests All in One. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining . 6095–6104

    2024

  24. [32]

    Yantao Yu, Weipeng Wang, Zhoutian Feng, and Daiyue Xue. 2021. A dual aug- mented two-tower model for online large-scale recommendation. DLP-KDD (2021)

    2021

  25. [33]

    Moyu Zhang, Yongxiang Tang, Jinxin Hu, and Yu Zhang. 2024. Scenario-adaptive fine-grained personalization network: Tailoring user behavior representation to the scenario context. In International ACM SIGIR Conference on Research and Development in Information Retrieval . 1557–1566

    2024

  26. [34]

    Yuanliang Zhang, Xiaofeng Wang, Jinxin Hu, Ke Gao, Chenyi Lei, and Fei Fang

  27. [35]

    Pengyu Zhao, Xin Gao, Chunxu Xu, and Liang Chen. 2023. M5: Multi-modal multi-interest multi-scenario matching for over-the-top recommendation. In ACM SIGKDD International Conference on Knowledge Discovery & Data Mining . 5650–5659

    2023

  28. [36]

    Guorui Zhou, Xiaoqiang Zhu, Chenru Song, Ying Fan, Han Zhu, Xiao Ma, Yanghui Yan, Junqi Jin, Han Li, and Kun Gai. 2018. Deep interest network for click-through rate prediction. In ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. 1059–1068

    2018

  29. [37]

    Jie Zhou, Xianshuai Cao, Wenhao Li, Lin Bo, Kun Zhang, Chuan Luo, and Qian Yu. 2023. HiNet: Novel multi-scenario & multi-task learning with hierarchical information extraction. In International Conference on Data Engineering . 2969– 2975

    2023

  30. [38]

    Feng Zhu, Yan Wang, Chaochao Chen, Jun Zhou, Longfei Li, and Guanfeng Liu

  31. [39]

    Jiachen Zhu, Yichao Wang, Jianghao Lin, Jiarui Qin, Ruiming Tang, Weinan Zhang, and Yong Yu. 2024. M-scan: A multi-scenario causal-driven adaptive network for recommendation. In The Web Conference. 3844–3853. PERSCEN: Learning Personalized Interaction Pattern and Scenario Pref...

    2024

  32. [42]

    In International Joint Conference on Artificial Intelligence

    Cross-domain recommendation: Challenges, progress, and prospects. In International Joint Conference on Artificial Intelligence . 4721–4728

  33. [2021]

    In ACM International Conference on Information & Knowledge Management

    SAR-Net: A scenario-aware ranking network for personalized fair recom- mendation in hundreds of travel scenarios. In ACM International Conference on Information & Knowledge Management . 4094–4103

  34. [2022]

    In ACM International Conference on Information & Knowl- edge Management

    Scenario-adaptive and self-supervised model for multi-scenario personal- ized recommendation. In ACM International Conference on Information & Knowl- edge Management. 3674–3683

  35. [2024]

    In ACM SIGKDD International Conference on Knowledge Discovery & Data Mining

    Warming up cold-start CTR prediction by learning item-specific feature interactions. In ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. 3233–3244

Pith tools

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