Pith. sign in

REVIEW 4 major objections 4 minor 62 references

Combinatorial Optimization Perspective based Framework for Multi-behavior Recommendation

T0 review · 4 major / 4 minor · reviewed 2026-08-09 · deepseek-v4-flash

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

Pith's one-line read Multi-behavior recommendation improves when fusion is treated as a constrained combinatorial search and prediction decouples feature and label signals.

desk verdict A solid incremental multi-behavior recommendation system with large reported gains, but the evaluation protocol is underspecified enough that the headline numbers need scrutiny before they carry weight. read the letter →

arxiv 2502.02232 v1 pith:TU2UTXDM submitted 2025-02-04 cs.IR

Chenhao Zhai , Chang Meng , Yu Yang , Kexin Zhang , Xuhao Zhao , Xiu Li This is my paper · ORCID
classification cs.IR
keywords combinatorialoptimizationmulti-behaviorrecommendationmulti-tasklearninggraphconvolutionalnetworknegativetransfercontrastiverecommendersystems
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 tries to establish that multi-behavior recommendation should be built as a two-step pipeline with explicit structure: a fusion step that treats user behavior patterns as a combinatorial search space and prunes it with staged constraints, and a prediction step that coordinates target and auxiliary tasks so that feature and label distribution gaps do not cause negative transfer. It proposes COPF, made of COGCN for fusion and DFME for prediction, and reports Hit Ratio@10 and NDCG@10 on Beibei, Taobao, and Tmall, with relative HR improvements of 49.91%, 12.06%, and 24.12% over the best baseline on each dataset. A sympathetic reading of the paper is that its staged constraints—pre-behavior, in-behavior, and post-behavior—capture user behavior patterns better than both unconstrained aggregation and strict cascading paradigms, and that its decoupling and stop-gradient designs make auxiliary tasks useful rather than harmful. If correct, the framework gives a reusable recipe for squeezing signal out of sparse, multi-behavior interaction data.

What carries the argument

The load-bearing mechanism is a staged constraint scheme on user behavior patterns, formalized in Definitions 3–5: a pre-behavior constraint channels only upstream behavior outputs into the current behavior's encoder, an in-behavior constraint restricts graph convolution on behavior k to adjacency matrices of behaviors 1...k (Equation 5), and a post-behavior constraint keeps each behavior's output separate for downstream tasks. The second mechanism is DFME, which coordinates the multi-task prediction head: contrastive loss (Equation 8), behavior-fitting experts built by mixing the current and another behavior's representations with small scaling coefficients (Equation 10), and a stop-gradient operation on the auxiliary-to-target aggregation path (Equation 14). Together they turn fusion into a pruned search and prevent auxiliary tasks from corrupting target-task gradients.

What would settle it

On a dataset with timestamps for every user-item behavior event, rebuild the behavior order per user from the actual event times and retrain COPF with the constraint order replaced by the observed order. If COPF's advantage over unconstrained or cascading baselines shrinks or disappears under the true order, then the reported gains are an artifact of the assumed view→cart→buy ordering rather than a general property of staged constraints.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is that the two weaknesses of current multi-behavior recommenders—imprecise fusion and uncoordinated multi-task prediction—can be fixed by imposing structure at the right points. COGCN formulates fusion as combinatorial optimization and restricts the solution space with three constraints: pre-behavior (the encoder of behavior k receives only outputs of upstream behaviors), in-behavior (message passing over the current behavior's adjacency matrix never uses semantic information from downstream behaviors), and post-behavior (decoupled outputs for each behavior feed separate prediction tasks). DFME then coordinates prediction: contrastive learning aligns target and auxiliary behavior representations, behavior-fitting experts use small mixing coefficients to refine the target's representation space, and a stop-gradient operation prevents auxiliary-task gradients from updating the target task. The paper argues, with ablations, that removing or relaxing any of these stages reduces performance, and that DFME plugs into other backbones and improves them.

Load-bearing premise

The framework assumes one fixed, globally shared order of behaviors (for instance, view before cart before buy) and that during fusion it is safe to exclude all downstream behavior information, even though real interaction timings are not observed.

Editorial extensions

If this is right

  • If COPF's constraints are right, multi-behavior recommenders do not have to choose between free-form aggregation and overly strict cascades; staged constraints are the better middle ground.
  • Auxiliary behaviors (views, carts, collects) can be used to improve buy prediction without the gradient interference that hurts many multi-task recommenders.
  • DFME should work as a drop-in prediction module: the compatibility study shows it raises accuracy when attached to prior fusion backbones such as LightGCN, MB-CGCN, CRGCN, CIGF, and PKEF.
  • The framework's success on three datasets with different behavior distributions suggests the gains are not tied to one platform's interaction patterns.

Reading between the lines

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

  • Editorial inference: the fixed global order assumption is the point to attack; the Appendix admits interaction timings are unknown, so a dataset with timestamps could re-order behaviors per user and reveal whether the staged constraints help or merely encode a prior.
  • Editorial inference: a natural extension is to learn a per-user or per-item probability distribution over behavior order instead of a single shared order, which would relax Definition 2.
  • Editorial inference: the contrastive-alignment plus stop-gradient pattern could be transferred to other multi-task ranking settings, such as multi-scenario prediction, where label distributions also differ across tasks.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 4 minor

Summary. The paper proposes COPF, a multi-behavior recommendation framework with two components: COGCN, which applies staged behavioral constraints (pre-behavior, in-behavior, post-behavior) to graph-convolutional message passing, and DFME, which combines contrastive alignment, behavior-fitting experts, and stop-gradient operators to mitigate negative transfer in multi-task prediction. On Beibei, Taobao, and Tmall, COPF reports HR@10 improvements of 49.91%, 12.06%, and 24.12% over the best baselines, with ablations and compatibility experiments supporting the contribution of each module. The authors release code at https://github.com/1918190/COPF.

Significance. If the empirical results are reproducible under a clean evaluation protocol, the reported gains are substantial and the ablation and compatibility analyses are unusually thorough for this area. The paper ships code, which is a concrete strength, and the modular design (COGCN and DFME can be studied separately) makes the claims testable. The main weakness is that the central claim is empirical while the evaluation protocol is under-specified; the combinatorial-optimization framing is also not formalized as an optimization problem. The paper does not provide machine-checked proofs or parameter-free derivations, so its value rests almost entirely on the reliability of the experiments.

major comments (4)
  1. [§5.1, Eqs. (5)–(6), (11)] The evaluation protocol is not specified: the paper does not state how the data are split into train, validation, and test sets, how negative items are sampled for BPR, or whether any hyperparameter selection uses the test set. Most importantly, it never states that test target interactions are removed from the behavior matrices B_k before training and before computing the graph convolutions in Eqs. (5), (6), and (11). Since COGCN and the behavior-fitting experts propagate over A_k, a test positive left in B_k would directly leak the answer into the representation computation. This is load-bearing for the SOTA claim in Table 2. Please specify the exact split (e.g., leave-one-out per user, temporal split, or random split), the negative sampling rule, the validation procedure, and confirm in the text that all test edges are excluded from all adjacency matrices used in training and in the forward pass at evaluation time.
  2. [§5.1.1, Table 2] Per-dataset hyperparameter searches are described only as ranges (e.g., GCN layers in {1,2,3,4}, behavior loss coefficients in {0,1/6,...,1}, scaling factors in {0.1,0.2,0.4,0.8}), but the final per-dataset values are not reported. The paper also reports averages over five runs without standard deviations and marks significance with a star without describing the statistical test. The magnitude of the claimed improvement, especially the 49.91% HR@10 gain on Beibei, cannot be assessed without knowing the selected configuration and the variance across runs. Please report the chosen hyperparameters for each dataset and the standard deviations, and describe the significance test used for the stars.
  3. [§4.2, Definitions 1–5] The 'combinatorial optimization perspective' is not formalized. The paper counts the number of possible user behavior patterns and states that constraints restrict the solution space, but it never defines the objective function, the decision variables, the feasible set, or the optimization problem being solved. The pre-, in-, and post-behavior constraints are architectural heuristics, and the claimed connection to combinatorial optimization is therefore not established by the text. Either provide a formal statement of the optimization problem with explicit constraints, or rephrase the contribution as a design rationale rather than a solution to a formal combinatorial problem.
  4. [Appendix A.2, Definition 2 and Eq. (5)] The method assumes a fixed, globally shared behavior order (view before cart before buy, or view before collect before cart before buy) while Appendix A.2 explicitly states that the exact timing of user-item interactions is unknown, so the assumed behavior patterns are not directly observed. If the assumed order is violated for a substantial fraction of users, the constraints in Definition 4 and Eq. (5) will discard useful signal. The paper should validate this assumption empirically, for example by comparing with a reversed behavior order, a variant that allows multiple orders, or by reporting the proportion of user histories that are consistent with the assumed order. Without such evidence, part of the reported gain could be an artifact of the ordering assumption rather than a general property of the method.
minor comments (4)
  1. [Eq. (11)] Equation (11) uses R^{k',l} with an initial state R^{k',1}, but R^{k',0} is never defined, and the dimensions and initialization of R^{k',1} are not stated. Please clarify how the hierarchical behavior embedding matrix is initialized and updated.
  2. [Figures 3–5] The parameter-analysis figures do not include error bars despite the paper reporting five runs, and some axis labels contain typos such as 'T aobao'. Adding error bars would make the claims about optimal layer counts and scaling factors more convincing.
  3. [§5.1.2, Table 1] The description says duplicate user-item interactions are eliminated by retaining the earliest one, but it is unclear whether this rule is applied per behavior type or across all behaviors, and how it affects the construction of B_k. Please clarify the preprocessing for each behavior matrix.
  4. [Abstract and §6] The module name is introduced as 'Distributed Fitting Multi-Expert Network' in the abstract and Section 4.3, but the conclusion calls it 'Distribution Fitting Expert Network'. Please use one consistent name throughout.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: COPF's central claims are benchmark-grounded empirical results, and each module's contribution is tested by ablations independent of the model's definitions.

full rationale

The paper's central claim is the empirical SOTA comparison in Table 2 against 15 external baselines on three public datasets; these improvements are not algebraic consequences of the model's own definitions. The COGCN constraints are design choices explicitly realized by Eqs. (3)-(7), and their value is tested by the ablations in Table 3 (COPF-P/A/F/D/C/B/H) and by the alternative aggregation and MTL module comparisons in Tables 4-5. The DFME mechanisms are also explicit objectives: the contrastive alignment Eq. (8), the behavior-fitting experts Eqs. (10)-(12), and the gate/stop-gradient rule Eq. (14) are all part of the training objective, and their contribution is ablated via w/o con., w/o for., w/o back., all sg., and w/o fit. No prediction reduces by construction to a fitted constant or to a renamed input. The self-citations to PKEF [31] and CIGF [15], which share authors, are used for conventions such as embedding size or decoupled input and as baselines; they are not invoked as an external uniqueness theorem or as the justification for the central claim. The reviewer-flagged lack of an explicit data split is a potential label-leakage or experimental-validity concern, not a definitional circularity, so it does not raise the circularity score under the stated rules. Overall, the derivation chain is self-contained: the method is defined, implemented, and then judged against external benchmarks and internal ablations.

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

No new physical entities are introduced. COGCN and DFME are network modules, not entities requiring independent evidence. The user behavior pattern is a conceptual definition tied to data, not a postulated object. The main ledger items are the tuned hyperparameters and the domain assumptions about behavior ordering and gradient decoupling.

free parameters (5)
  • Behavior loss coefficients lambda_k = Not reported per dataset; searched in {0,1/6,2/6,3/6,4/6,5/6,1} with sum fixed to 1
    Section 5.1.1: coefficients per behavior are tuned; final values are not given, so reported metrics may reflect favorable settings.
  • Number of GCN layers L = Not reported per dataset; searched in {1,2,3,4}
    Section 5.1.1 and Figure 3 show the optimal layer count varies by dataset; the final value per dataset is not stated.
  • Contrastive temperature tau = Not reported per dataset; searched in {0.1,0.2,0.3,0.4,0.5,0.6,0.8}
    Section 5.5.2 and Figure 4: temperature is selected by search, but no final tau per dataset is given.
  • Scaling coefficients alpha and beta = alpha searched in {0.1,0.2,0.4,0.8}; beta fixed at 0.001
    Section 4.3.2 and Figure 5: alpha controls the mixing in behavior-fitting experts and is tuned per dataset.
  • Contrastive loss coefficient gamma and L2 coefficient mu = gamma=1, mu=0.01
    Section 5.1.1: set by hand without justification; these are design constants.
assumptions (5)
  • domain assumption A single global behavior order exists and is known in advance (e.g., view, cart, buy).
    Definition 2 and the pre-behavior and in-behavior constraints in Section 4.2 assume this order; no per-user timing evidence is used.
  • domain assumption Downstream behavior information should be excluded from a behavior's node representation during fusion to prevent information leakage and overfitting.
    Definition 4 and Eq. (5): the in-behavior constraint sums only upstream adjacency matrices; this is a modeling choice, not a proven property.
  • ad hoc to paper Gradient updates from auxiliary tasks to the target behavior representation are harmful and should be blocked.
    Eq. (14) applies stop-gradient when j != k and j = K; the paper argues this reduces negative transfer but provides no theoretical guarantee.
  • ad hoc to paper The combinatorial optimization view is valid, i.e., the number of possible user behavior patterns is sum_{i=1}^{K} K!/(K-i)! and restricting the solution space by constraints is the right formalization.
    Section 4.2.1, Definitions 1-5: this is presented as a framing with no optimization objective or optimality proof, and the counting is not connected to the model's complexity.
  • ad hoc to paper The behavior-fitting expert update in Eq. (11) is well-defined.
    R^{k',l} and its initialization are not specified; W^l sharing with the GCN layer is inconsistent with the parameter-free update in Eq. (5).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Combinatorial Optimization Perspective based Framework for Multi-behavior Recommendation." pith.science (2026). https://pith.science/paper/TU2UTXDM

@misc{pith2026250202232,
  author       = {Pith},
  title        = {Pith review of: Combinatorial Optimization Perspective based Framework for Multi-behavior Recommendation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TU2UTXDM}},
  note         = {Machine review of arXiv:2502.02232}
}
read the original abstract

In real-world recommendation scenarios, users engage with items through various types of behaviors. Leveraging diversified user behavior information for learning can enhance the recommendation of target behaviors (e.g., buy), as demonstrated by recent multi-behavior methods. The mainstream multi-behavior recommendation framework consists of two steps: fusion and prediction. Recent approaches utilize graph neural networks for multi-behavior fusion and employ multi-task learning paradigms for joint optimization in the prediction step, achieving significant success. However, these methods have limited perspectives on multi-behavior fusion, which leads to inaccurate capture of user behavior patterns in the fusion step. Moreover, when using multi-task learning for prediction, the relationship between the target task and auxiliary tasks is not sufficiently coordinated, resulting in negative information transfer. To address these problems, we propose a novel multi-behavior recommendation framework based on the combinatorial optimization perspective, named COPF. Specifically, we treat multi-behavior fusion as a combinatorial optimization problem, imposing different constraints at various stages of each behavior to restrict the solution space, thus significantly enhancing fusion efficiency (COGCN). In the prediction step, we improve both forward and backward propagation during the generation and aggregation of multiple experts to mitigate negative transfer caused by differences in both feature and label distributions (DFME). Comprehensive experiments on three real-world datasets indicate the superiority of COPF. Further analyses also validate the effectiveness of the COGCN and DFME modules. Our code is available at https://github.com/1918190/COPF.

Figures

Figures reproduced from arXiv: 2502.02232 by the authors.

Figure 1
Figure 1. Examples of the user behavior patterns. traditional matrix factorization [23, 37, 41] to deep neural net￾works [10, 14, 47]. Among them, graph neural networks (GNNs) [5, 16, 32, 33, 43] have become popular techniques in multi-behavior recommendation due to their ability to model higher-order user￾item interactions effectively [4, 31, 46, 48]. For example, MBGCN [21] and CIGF [15] learn different behavioral informati… view at source ↗
Figure 2
Figure 2. Illustration of the proposed COPF framework, and we use three user behavior types as examples: view, cart, and buy. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 4
Figure 4. Impact of temperature hyperparameter. Beibei Taobao Tmall 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 HR@10 α = 0.1 α = 0.2 α = 0.4 α = 0.8 Beibei Taobao Tmall 0.00 0.02 0.04 0.06 0.08 0.10 NDCG@10 α = 0.1 α = 0.2 α = 0.4 α = 0.8 [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figures from the paper (4 more)
Figure 5
Figure 5. Figure 5: Impact of the scaling size. 5.5.3 Impact of the scaling size. With 𝛽 fixed at 0.001, we investi￾gate the impact of the scaling factor 𝛼 on the forward propagation in multi-task learning [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: shows the data distribution on the three datasets. 1/0 indicates the presence or absence of the corresponding type of behavior. For example, 1001 represents users who only have view and buy behaviors with items. As we can see, there are significant differences in the b…
Figure 7
Figure 7. Figure 7: The structure of transfer-based MTL. Red line rep [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: Refine with other behavior representions. [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

62 extracted references · 41 canonical work pages

  1. [1]

    Martín Abadi, Paul Barham, Jianmin Chen, Zhifeng Chen, Andy Davis, Jeffrey Dean, Matthieu Devin, Sanjay Ghemawat, Geoffrey Irving, Michael Isard, et al

  2. [2]

    Zhi Bian, Shaojun Zhou, Hao Fu, Qihong Yang, Zhenqi Sun, Junjie Tang, Guiquan Liu, Kaikui Liu, and Xiaolong Li. 2021. Denoising user-aware memory network for recommendation. In Fifteenth ACM Conference on Recommender Systems . 400–410

    work page 2021

  3. [3]

    Rich Caruana. 1997. Multitask learning. Machine learning 28, 1 (1997), 41–75

    1997

  4. [4]

    Chong Chen, Weizhi Ma, Min Zhang, Zhaowei Wang, Xiuqiang He, Chenyang Wang, Yiqun Liu, and Shaoping Ma. 2021. Graph Heterogeneous Multi-Relational Recommendation. In AAAI, Vol. 35. 3958–3966

    work page 2021

  5. [5]

    Lei Chen, Le Wu, Richang Hong, Kun Zhang, and Meng Wang. 2020. Revisiting Graph Based Collaborative Filtering: A Linear Residual Graph Convolutional Network Approach. In AAAI, Vol. 34. 27–34

    work page 2020

  6. [6]

    Zhiyong Cheng, Sai Han, Fan Liu, Lei Zhu, Zan Gao, and Yuxin Peng. 2023. Multi-Behavior Recommendation with Cascading Graph Convolution Networks. In Proceedings of the ACM Web Conference 2023 . 1181–1189

    work page 2023

  7. [7]

    Yizhou Dang, Enneng Yang, Guibing Guo, Linying Jiang, Xingwei Wang, Xiaoxiao Xu, Qinghui Sun, and Hong Liu. 2023. Uniform sequence better: Time interval aware data augmentation for sequential recommendation. In Proceedings of the AAAI conference on artificial intelligence , Vol. 37. 4225–4232

    work page 2023

  8. [8]

    Yizhou Dang, Jiahui Zhang, Yuting Liu, Enneng Yang, Yuliang Liang, Guibing Guo, Jianzhe Zhao, and Xingwei Wang. 2024. Augmenting Sequential Recommendation with Balanced Relevance and Diversity. arXiv preprint arXiv:2412.08300 (2024)

    arXiv 2024

Show all 62 references
  1. [9]

    Jingtao Ding, Guanghui Yu, Xiangnan He, Yuhan Quan, Yong Li, Tat-Seng Chua, Depeng Jin, and Jiajie Yu. 2018. Improving Implicit Recommender Systems with View Data.. In IJCAI. 3343–3349

    2018

  2. [10]

    Chen Gao, Xiangnan He, Dahua Gan, Xiangning Chen, Fuli Feng, Yong Li, Tat- Seng Chua, and Depeng Jin. 2019. Neural multi-task recommendation from multi-behavior data. In ICDE. IEEE, 1554–1557

    2019

  3. [11]

    Xavier Glorot and Yoshua Bengio. 2010. Understanding the difficulty of training deep feedforward neural networks. In AISTATS. 249–256

    2010

  4. [12]

    Shuyun Gu, Xiao Wang, Chuan Shi, and Ding Xiao. 2022. Self-supervised Graph Neural Networks for Multi-behavior Recommendation. In Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence, IJCAI-22. 2052– 2058

    2022

  5. [13]

    Guibing Guo, Huihuai Qiu, Zhenhua Tan, Yuan Liu, Jing Ma, and Xingwei Wang

  6. [14]

    Long Guo, Lifeng Hua, Rongfei Jia, Binqiang Zhao, Xiaobo Wang, and Bin Cui. 2019. Buying or browsing?: Predicting real-time purchasing intent using attention-based deep network with multiple behavior. In SIGKDD. 1984–1992

    2019

  7. [15]

    Wei Guo, Chang Meng, Enming Yuan, Zhicheng He, Huifeng Guo, Yingxue Zhang, Bo Chen, Yaochen Hu, Ruiming Tang, Xiu Li, et al. 2023. Compressed Interaction Graph based Framework for Multi-behavior Recommendation. In Proceedings of the ACM Web Conference 2023. 960–970

    2023

  8. [16]

    Xiangnan He, Kuan Deng, Xiang Wang, Yan Li, Yongdong Zhang, and Meng Wang. 2020. LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation. arXiv preprint arXiv:2002.02126 (2020)

    2020 arXiv

  9. [17]

    Xiangnan He, Lizi Liao, Hanwang Zhang, Liqiang Nie, Xia Hu, and Tat-Seng Chua. 2017. Neural collaborative filtering. In WWW. 173–182

    2017

  10. [18]

    Zhicheng He, Weiwen Liu, Wei Guo, Jiarui Qin, Yingxue Zhang, Yaochen Hu, and Ruiming Tang. 2023. A Survey on User Behavior Modeling in Recommender Systems. arXiv preprint arXiv:2302.11087 (2023)

    2023 arXiv

  11. [19]

    Chao Huang. 2021. Recent Advances in Heterogeneous Relation Learning for Recommendation. arXiv preprint arXiv:2110.03455 (2021)

    2021 arXiv

  12. [20]

    Robert A Jacobs, Michael I Jordan, Steven J Nowlan, and Geoffrey E Hinton. 1991. Adaptive mixtures of local experts. Neural computation 3, 1 (1991), 79–87

    1991

  13. [21]

    Bowen Jin, Chen Gao, Xiangnan He, Depeng Jin, and Yong Li. 2020. Multi- behavior recommendation with graph convolutional networks. In SIGIR

    2020

  14. [22]

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

    2014 arXiv

  15. [23]

    Artus Krohn-Grimberghe, Lucas Drumond, Christoph Freudenthaler, and Lars Schmidt-Thieme. 2012. Multi-relational matrix factorization using bayesian personalized ranking for social network data. In Proceedings of the fifth ACM international conference on Web search and data min...

    2012

  16. [24]

    Pengcheng Li, Runze Li, Qing Da, An-Xiang Zeng, and Lijun Zhang. 2020. Improv- ing multi-scenario learning to rank in e-commerce by exploiting task relation- ships in the label space. In Proceedings of the 29th ACM International Conference on Information & Knowledge Management...

    2020

  17. [25]

    Xiang Li, Chaofan Fu, Zhongying Zhao, Guanjie Zheng, Chao Huang, Junyu Dong, and Yanwei Yu. 2024. Dual-Channel Multiplex Graph Neural Networks for Recommendation. arXiv preprint arXiv:2403.11624 (2024)

    2024 arXiv

  18. [26]

    Xiaodong Li, Jiawei Sheng, Jiangxia Cao, Wenyuan Zhang, Quangang Li, and Tingwen Liu. 2024. CDRNP: Cross-Domain Recommendation to Cold-Start Users via Neural Process. In Proceedings of the 17th ACM International Conference on Web Search and Data Mining. 378–386

    2024

  19. [27]

    Ke Liang, Yue Liu, Sihang Zhou, Wenxuan Tu, Yi Wen, Xihong Yang, Xiangjun Dong, and Xinwang Liu. 2023. Knowledge Graph Contrastive Learning Based on Relation-Symmetrical Structure. IEEE Transactions on Knowledge and Data Engineering (2023)

    2023

  20. [28]

    Babak Loni, Roberto Pagano, Martha Larson, and Alan Hanjalic. 2016. Bayesian personalized ranking with multi-channel user feedback. In Proceedings of the 10th ACM Conference on Recommender Systems . 361–364

    2016

  21. [29]

    Jiaqi Ma, Zhe Zhao, Xinyang Yi, Jilin Chen, Lichan Hong, and Ed H Chi. 2018. Modeling task relationships in multi-task learning with multi-gate mixture-of- experts. In SIGKDD. 1930–1939

    2018

  22. [30]

    Chang Meng, Chenhao Zhai, Xueliang Wang, Shuchang Liu, Xiaoqiang Feng, Lantao Hu, Xiu Li, Han Li, and Kun Gai. 2024. Coarse-to-fine Dynamic Uplift Modeling for Real-time Video Recommendation. arXiv preprint arXiv:2410.16755 (2024)

    2024 arXiv

  23. [31]

    Chang Meng, Chenhao Zhai, Yu Yang, Hengyu Zhang, and Xiu Li. 2023. Parallel Knowledge Enhancement based Framework for Multi-behavior Recommendation. In Proceedings of the 32nd ACM International Conference on Information and Knowledge Management. 1797–1806

    2023

  24. [32]

    Chang Meng, Hengyu Zhang, Wei Guo, Huifeng Guo, Haotian Liu, Yingxue Zhang, Hongkun Zheng, Ruiming Tang, Xiu Li, and Rui Zhang. 2023. Hierarchical Projection Enhanced Multi-Behavior Recommendation. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Dat...

    2023

  25. [33]

    Chang Meng, Ziqi Zhao, Wei Guo, Yingxue Zhang, Haolun Wu, Chen Gao, Dong Li, Xiu Li, and Ruiming Tang. 2022. Coarse-to-Fine Knowledge-Enhanced Multi-Interest Learning Framework for Multi-Behavior Recommendation. arXiv preprint arXiv:2208.01849 (2022)

    2022 arXiv

  26. [34]

    Huihuai Qiu, Yun Liu, Guibing Guo, Zhu Sun, Jie Zhang, and Hai Thanh Nguyen

  27. [35]

    Steffen Rendle, Christoph Freudenthaler, Zeno Gantner, and Lars Schmidt-Thieme

  28. [36]

    Michael Schlichtkrull, Thomas N Kipf, Peter Bloem, Rianne Van Den Berg, Ivan Titov, and Max Welling. 2018. Modeling relational data with graph convolutional networks. In European semantic web conference . Springer, 593–607

    2018

  29. [37]

    Ajit P Singh and Geoffrey J Gordon. 2008. Relational learning via collective matrix factorization. In Proceedings of the 14th ACM SIGKDD international conference on Knowledge discovery and data mining . 650–658

    2008

  30. [38]

    Liangcai Su, Junwei Pan, Ximei Wang, Xi Xiao, Shijie Quan, Xihua Chen, and Jie Jiang. 2024. STEM: Unleashing the Power of Embeddings for Multi-task Recommendation. In Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 38. 9002–9010

    2024

  31. [39]

    Xiaoyuan Su and Taghi M Khoshgoftaar. 2009. A survey of collaborative filtering techniques. Advances in artificial intelligence 2009 (2009)

    2009

  32. [40]

    Hongyan Tang, Junning Liu, Ming Zhao, and Xudong Gong. 2020. Progressive layered extraction (ple): A novel multi-task learning (mtl) model for personalized recommendations. In RecSys. 269–278

    2020

  33. [41]

    Liang Tang, Bo Long, Bee-Chung Chen, and Deepak Agarwal. 2016. An empirical study on recommendation with multiple types of feedback. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. 283–292

    2016

  34. [42]

    Lisa Torrey and Jude Shavlik. 2010. Transfer learning. In Handbook of research on machine learning applications and trends: algorithms, methods, and techniques . IGI global, 242–264

    2010

  35. [43]

    Xiang Wang, Xiangnan He, Meng Wang, Fuli Feng, and Tat-Seng Chua. 2019. Neural graph collaborative filtering. In SIGIR. 165–174

    2019

  36. [44]

    Xiaobei Wang, Shuchang Liu, Xueliang Wang, Qingpeng Cai, Lantao Hu, Han Li, Peng Jiang, Kun Gai, and Guangming Xie. 2024. Future Impact Decomposition in Request-level Recommendations. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining . 5905–5916

    2024

  37. [45]

    Wei Wei, Chao Huang, Lianghao Xia, Yong Xu, Jiashu Zhao, and Dawei Yin. 2022. Contrastive meta learning with behavior multiplicity for recommendation. In Proceedings of the Fifteenth ACM International Conference on Web Search and Data Mining. 1120–1128

    2022

  38. [46]

    Lianghao Xia, Chao Huang, Yong Xu, Peng Dai, Mengyin Lu, and Liefeng Bo. 2021. Multi-Behavior Enhanced Recommendation with Cross-Interaction Collaborative Relation Modeling. In ICDE. IEEE, 1931–1936

    2021

  39. [47]

    Lianghao Xia, Chao Huang, Yong Xu, Peng Dai, Bo Zhang, and Liefeng Bo

  40. [48]

    Lianghao Xia, Chao Huang, Yong Xu, Peng Dai, Xiyue Zhang, Hongsheng Yang, Jian Pei, and Liefeng Bo. 2021. Knowledge-enhanced hierarchical graph trans- former network for multi-behavior recommendation. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 35. 4486–4493

    2021

  41. [49]

    Lianghao Xia, Yong Xu, Chao Huang, Peng Dai, and Liefeng Bo. 2021. Graph meta network for multi-behavior recommendation. In SIGIR. 757–766. Combinatorial Optimization Perspective based Framework for Multi-behavior Recommendation KDD ’25, August 3–7, 2025, Toronto, ON, Canada

    2021

  42. [50]

    Hongrui Xuan, Yi Liu, Bohan Li, and Hongzhi Yin. 2023. Knowledge Enhancement for Contrastive Multi-Behavior Recommendation. In Proceedings of the Sixteenth ACM International Conference on Web Search and Data Mining . 195–203

    2023

  43. [51]

    Mingshi Yan, Zhiyong Cheng, Chen Gao, Jing Sun, Fan Liu, Fuming Sun, and Haojie Li. 2022. Cascading Residual Graph Convolutional Network for Multi- Behavior Recommendation. arXiv preprint arXiv:2205.13128 (2022)

    2022 arXiv

  44. [52]

    Mingshi Yan, Fan Liu, Jing Sun, Fuming Sun, Zhiyong Cheng, and Yahong Han

  45. [53]

    Chengqing Yu, Guangxi Yan, Chengming Yu, Yu Zhang, and Xiwei Mi. 2023. A multi-factor driven spatiotemporal wind power prediction model based on ensemble deep graph attention reinforcement learning networks. Energy 263 (2023), 126034

    2023

  46. [54]

    Xuhao Zhao, Yanmin Zhu, Chunyang Wang, Mengyuan Jing, Jiadi Yu, and Feilong Tang. 2023. Task-difficulty-aware meta-learning with adaptive update strategies for user cold-start recommendation. In Proceedings of the 32nd ACM International Conference on Information and Knowledge ...

    2023

  47. [55]

    Zhe Zhao, Zhiyuan Cheng, Lichan Hong, and Ed H Chi. 2015. Improving user topic interest profiles by behavior factorization. In Proceedings of the 24th International Conference on World Wide Web. 1406–1416

    2015

  48. [56]

    cart" and

    Chang Zhou, Jinze Bai, Junshuai Song, Xiaofei Liu, Zhengchao Zhao, Xiusi Chen, and Jun Gao. 2018. Atrank: An attention-based user behavior modeling frame- work for recommendation. In Thirty-Second AAAI Conference on Artificial Intelli- gence. A APPENDIX A.1 Complexity Analysis...

    2018

  49. [2012]

    arXiv preprint arXiv:1205.2618 (2012)

    BPR: Bayesian personalized ranking from implicit feedback. arXiv preprint arXiv:1205.2618 (2012)

    2012 arXiv

  50. [2016]

    In 12th USENIX symposium on operating systems design and implementation (OSDI 16)

    {TensorFlow}: a system for {Large-Scale} machine learning. In 12th USENIX symposium on operating systems design and implementation (OSDI 16) . 265–283

  51. [2017]

    Knowledge-Based Systems 138 (2017), 202–207

    Resolving data sparsity by multi-type auxiliary implicit feedback for recommender systems. Knowledge-Based Systems 138 (2017), 202–207

    2017

  52. [2018]

    Information Sciences 453 (2018), 80–98

    BPRH: Bayesian personalized ranking for heterogeneous implicit feedback. Information Sciences 453 (2018), 80–98

    2018

  53. [2020]

    In SIGIR

    Multiplex behavioral relation learning for recommendation via memory augmented transformer network. In SIGIR. 2397–2406

  54. [2024]

    In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval

    Behavior-Contextualized Item Preference Modeling for Multi-Behavior Recommendation. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval . 946–955

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.