Behavior Modeling Space Reconstruction for E-Commerce Search

Chi Zhang; Ling Zhong; Maolin Wang; Qidong Liu; Wei Lin; Xiangyu Zhao; Xing Shi; Xudong Yang; Xuetao Wei; Yejing Wang

REVIEW 3 major objections 6 minor 66 references

Behavior Modeling Space Reconstruction for E-Commerce Search

T0 review · 3 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read DRP improves e-commerce search prediction by editing preference representations to remove relevance's influence and fusing the two signals adaptively, without labeled relevance data.

desk verdict Solid adaptive fusion gains for e-commerce search, but the causal disentanglement claim is not verified by the paper's own experiments. read the letter →

arxiv 2501.18216 v3 pith:S4LUPBU5 submitted 2025-01-30 cs.IR

Yejing Wang , Chi Zhang , Xiangyu Zhao , Qidong Liu , Maolin Wang , Xuetao Wei , Zitao Liu , Xing Shi

show 3 more authors

Xudong Yang Ling Zhong Wei Lin

This is my paper · ORCID

classification cs.IR

keywords userbehaviormodelinge-commercesearchrelevanceandpreferencedisentanglementeditingadaptivefusionclick-throughratepredictioncausalgraphspace

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 behavior models in e-commerce search—which predict clicks by combining query–item relevance with user preference—are trained in a collapsed modeling space: relevance leaks into preference estimates, and a fixed multiplicative fusion formula cannot represent samples where relevance and preference disagree. To fix this, it introduces DRP, which edits the preference model's last-layer representation by subtracting a learned low-rank projection of the relevance representation, producing a calibrated preference score, and then combines relevance and preference with a dual-level adaptive fusion. If DRP is correct, search accuracy improves on click and ranking metrics without any human-labeled relevance judgments, which would make the approach much easier to deploy across e-commerce platforms. The claim is supported by consistent AUC and NDCG gains over the reported baselines on KuaiSAR, JDSearch, and a private dataset.

What carries the argument

The load-bearing device is the orthogonal low-rank projection matrix $\boldsymbol{O}\in\mathbb{R}^{H\times D}$. Because $\boldsymbol{O}$ is orthogonal, $\boldsymbol{O}^T$ is its inverse, so subtracting $\boldsymbol{O}\mathcal{e}_r$ inside the subspace and projecting back edits the preference representation exactly along the direction that best matches the causal influence of relevance on preference while preserving other information. The second device is the adaptive fusion formula, which rewrites the static product as a $2\times 2$ probability table $P_{ij}$ for each relevance–preference status and then applies learnable vectors $\boldsymbol{\alpha},\boldsymbol{\beta}$ for global adaptation and a residual function $F$ for local adaptation, separating samples that were previously collapsed, such as irrelevant-but-clicked items.

What would settle it

A reader could test the causal claim by constructing a dataset with known relevance labels and a planted relevance-to-preference influence, then checking whether the learned $\boldsymbol{O}$ recovers the planted direction and whether edited preference scores stop shifting when relevance changes but preference is fixed; either failure would falsify the central claim.

Watch

Extended reading notes

Core claim

The central discovery is that the causal graph for e-commerce behavior has a path $R\rightarrow P\rightarrow B$, not only $P\rightarrow B$ and $R\rightarrow B$, and that existing joint models ignore it. Treating the last-layer representations $\mathcal{e}_p$ and $\mathcal{e}_r$ as neural counterparts of preference and relevance, DRP learns an orthogonal low-rank matrix $\boldsymbol{O}$ whose row space approximates the relevance-to-preference intervention space, and computes $\mathcal{e}_{p_c} = \boldsymbol{O}^T(\boldsymbol{O}\mathcal{e}_p - \boldsymbol{O}\mathcal{e}_r)$. Then adaptive fusion replaces the fixed $\hat{r}^{\delta}\hat{p}$ product with learnable global and local coefficients so the model can distinguish all six areas of the Venn diagram. The paper claims this yields disentangled, untainted preference predictions and a reconstructed modeling space, with no auxiliary supervision beyond click labels.

Load-bearing premise

The load-bearing premise is that the learned orthogonal low-rank matrix $\boldsymbol{O}$ captures the true causal influence of relevance on preference, so subtracting $\boldsymbol{O}\mathcal{e}_r$ inside that subspace really removes the indirect relevance effect rather than merely reweighting the two representations.

Editorial extensions

If this is right

Click and ranking prediction in e-commerce search can be improved by removing the relevance-induced bias in preference representations while keeping the same backbone encoders.
The two components—preference editing and adaptive fusion—work across relevance backbones such as DSSM, QEM, and HEM and preference backbones such as MLP and DCN, so the improvement is not tied to one architecture.
Joint search models can be trained end-to-end from behavior signals alone, eliminating the need for expensive human relevance labels during training.
Samples where relevance and preference disagree, such as irrelevant-but-clicked items, receive distinct predictions instead of being collapsed into one score.

Reading between the lines

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

The same orthogonal-editing trick could decompose other confounded effects in search, such as position bias or exposure bias, wherever a causal path between two predictive signals exists.
If the learned $\boldsymbol{O}$ truly isolates the relevance-to-preference influence, DRP's edited preference scores should be stable across queries that vary only in relevance; this is a testable prediction the paper does not run.
Since no labeled relevance data is used, the approach may transfer to smaller platforms, but only if behavior signal alone is rich enough to identify the intervention subspace; the paper's experiments do not directly verify this.

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.

Desk Editor's Note

read the letter

Here's the short version. DRP is a solid engineering contribution to e-commerce click prediction, and the adaptive fusion component is the real workhorse. The paper's headline claim—that preference editing removes the relevance effect to give 'untainted' preferences—is not verified, and the ablation data suggest the editing itself contributes almost nothing to the reported gains.

What's actually new: combining a Geiger-style low-rank causal edit on the preference representation with a dual-level adaptive fusion, all trained without human-labeled relevance data. The identity in Equation (8) is a neat reframing of the fixed fusion rule, and the six-area Venn diagram is a clear way to talk about what joint models miss. The experimentation is extensive: three datasets (two public, one private), three relevance backbones, two preference backbones, several joint-modeling baselines, and an extra InvPref comparison in the appendix. The gains are modest but consistent, and significance tests are reported.

The soft spot is load-bearing. O in Equation (6) is learned only through the behavior loss in Equation (11). Nothing forces O to align with the true R→P intervention space. Geiger and collaborators identify intervention spaces using known causal variables or supervised alignments; neither is available here, so the edited representation is not demonstrated to isolate preference from relevance. The paper's own Table 4 makes this concrete: on JDSearch/MLP, preference editing alone (DRP-2) moves AUC from 0.6669 to 0.6671, while adaptive fusion alone (DRP-5) gives 0.6813, close to the full model's 0.6824. That is not a story about causal disentanglement driving the improvement. The visualization in Section 4.5 is also circular: areas are defined by the model's own top-20% relevance and preference scores, and then the model is shown to separate them. One more mild issue: the local fusion function F is under-specified—it is said to take only u, v, q during the feed-forward loop while using y_g and y to build its training objective, and it isn't clear what F actually computes at serving time.

Who should read this: people building joint relevance-preference models for product search. The adaptive fusion component is worth borrowing, and the framing is legible. The causal editing idea should be treated as a plausible hypothesis until it gets a direct test—for example, synthetic data with known interventions, or a held-out set of human relevance judgments. As an editor, I would send this to review: the experiments are honest and reproducible in spirit (code link is given, though no commit hash), and the central claim is precisely the kind of thing referees should pressure-test.

Referee Report

3 major / 6 minor

Summary. The paper studies e-commerce search behavior modeling, arguing that existing joint relevance/preference frameworks suffer from two problems: entangled relevance and preference effects (via the R->P causal path) and a collapsed modeling space caused by static fusion rules. It proposes DRP, which (i) edits the preference representation by subtracting a low-rank projection of the relevance representation (preference editing, Eq. 6) and (ii) replaces the fixed product fusion with a two-level adaptive fusion score (Eqs. 9-10). The method is evaluated on KuaiSAR, JDSearch, and a private dataset, under several preference and relevance backbones, reporting AUC, LogLoss, NDCG, and HR gains over joint-modeling baselines. The paper also provides a Venn-diagram analysis that divides the behavior space into six areas and argues that existing methods collapse these areas.

Significance. If the causal disentanglement claim were established, the paper would make a useful contribution: it offers a unified causal/Venn perspective on e-commerce search behavior, proposes a simple representation-editing mechanism that avoids human-labeled relevance data, and backs the proposal with unusually extensive experiments across three datasets and multiple backbone combinations. The derivation in Eq. (8) is correct, and the adaptive-fusion component appears to be a genuinely effective and practical contribution. However, the central claim that preference editing removes the relevance effect and yields 'untainted' preferences is not directly verified by the training objective or by the ablation results; as presented, the empirical gains are largely attributable to adaptive fusion rather than to the causal editing mechanism. The manuscript's contribution is therefore only partially supported in its current form.

major comments (3)

[§3.2, §3.4, Eq. (6), Eq. (11)] The causal claim that Equation (6) removes the relevance effect from preference predictions is not verified. The orthogonal matrix O is trained only with the behavior prediction loss (Eq. 11); nothing in that objective ties the learned low-rank subspace to the causal intervention space of R->P. The cited basis, Geiger et al. [13], identifies intervention spaces using known causal variables or supervised alignments, which are not available in this setting. Consequently, e_pc = O^T(Oe_p - Oe_r) is a behavior-prediction-optimized projection, and the interpretation of it as 'untainted preference' remains an assumption. Since the abstract and contributions state this as a core mechanism, the paper needs a direct verification, for example in a synthetic setting with a known R->P intervention or via an external relevance/preference ground truth on at least one dataset.
[§4.3, Table 4] The ablation undermines the causal disentanglement claim. On JDSearch with MLP, DRP-2 (preference editing without adaptive fusion) yields AUC 0.6671 versus Base 0.6669, a gain of only 0.0002, while DRP-5 (adaptive fusion without preference editing) yields 0.6813, a gain of 0.0144; the full DRP yields 0.6824. Thus nearly all improvement comes from adaptive fusion, and the preference-editing module is not shown to deliver the claimed disentanglement benefit. Please also report DRP-1 (removal of the orthogonal constraint) in Table 4, since the variant list defines it but the table omits it, and provide a check that the edited representation actually reduces the relevance effect.
[§4.5, Figure 4] The model-visualization section defines Area#0-5 using the model's own top-20% relevance and preference scores, then uses the model's own predictions to demonstrate separation of these areas. This is circular as evidence of disentanglement: if the relevance and preference scores are entangled, the area labels are not ground truth. The visualization can illustrate the behavior of the fusion module, but it cannot substantiate the claim that preference editing removes the relevance effect. An external or synthetic ground truth for relevance/preference status, or at least a robustness check with different area-definition thresholds, is needed.

minor comments (6)

[§3.3, Eq. (8)] The matrix in Eq. (8) is interpreted as the joint probabilities for the six relevance-preference areas, but this implicitly assumes independence of p_c and r and is not normalized after the learnable alpha/beta replacement in Eq. (9). The paper should state this assumption explicitly.
[§4.5] The sentence 'we incorporate an additional Area#0 following Area#6' is confusing: the paper defines six areas (Area#0-5), but the text refers to Area#6. Please clarify the labeling of the heatmap columns.
[§4.5] The text says 'The fixed fusion in Equation 7 fails to differentiate Area#2&4 and Area#1&5,' but Equation (7) defines p_c, not the final fusion score. The reference should be to Equation (5) or (8).
[§4.1.4] The initial values of alpha and beta, (1, 0.5), and the low-rank dimension D=16 are introduced as implementation details, but their selection is only tested on KuaiSAR. Please state how these hyperparameters were chosen and whether they transfer across datasets.
[Table 3] Several DRP entries do not carry asterisks (for example, KuaiSAR LogLoss with the DSSM row and JDSearch LogLoss with the DSSM row), yet the text claims DRP outperforms the alternatives. Please clarify which differences are statistically significant and whether the claim refers to all metrics or only the starred ones.
[§3.2] The notation for the low-rank projection matrix is inconsistent: the text defines O in Eq. (6) but later refers to 'D = 16 for R'. Please standardize the symbol.

Circularity Check

1 steps flagged · score 2.0 of 10

No load-bearing circularity; one self-referential visualization does not by itself validate the disentanglement claim.

self definitional [Section 4.5, Model Visualization]
"We classify samples with the top 20% relevance scores as relevant and the same for preference to construct areas. ... Through global adaptive fusion, DRP successfully discriminates among four distinct areas: Area#0, Area#1&2, Area#3, and Area#4&5 ... These findings are in accord with our theoretical analysis in Section 3.3."

The six 'areas' are constructed from the model's own top-20% relevance and preference scores. Since the behavior prediction y-hat is a function of those same scores (Equations 9-10), partitioning samples by those scores and then reporting different average y-hat per area is true by construction. The heatmap cannot confirm that the areas correspond to true relevance/preference statuses or that Equation (6) yields 'untainted' preferences; the validation is defined by the model's own outputs.

full rationale

The paper's headline improvements are external benchmark results (AUC, LogLoss, NDCG, HR) on KuaiSAR, JDSearch, and a private dataset, consistently beating Base, CLK, NISE, DCMT, and PRINT. These gains are not forced by the construction; they are genuine empirical comparisons. The main weakness is the causal interpretation of preference editing: Equation (6) learns the orthogonal matrix O using only the behavior loss in Equation (11), with no auxiliary signal or supervised alignment to guarantee that O spans the true R->P intervention space; the citation to Geiger et al. [13] supplies a general identification result, not a guarantee for this unsupervised setting. This is an evidential gap and a correctness risk, but it is not a circular reduction, because 'untainted preference' is not defined as 'whatever improves the behavior loss.' The single by-construction element is the Section 4.5 visualization, which defines areas using the model's own relevance/preference scores and then shows the model separates them; that is a self-referential check and is not load-bearing for the externally measured ranking gains. Consistent with hard rule #3, the presence of independent external benchmarks keeps the circularity score low.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The framework relies on a causal graph (R->P) and a low-rank intervention space that are asserted but not directly verified. The training loss is the standard behavior prediction loss, so the disentanglement is an interpretation of what the orthogonal projection does, not a measured property. The only free parameters are hyperparameters; no global constants are fitted to the final results.

free parameters (4)

delta (δ) = not reported
Hyperparameter balancing relevance and preference in the final scoring equation (Equations 3 and 5). No value or sensitivity analysis is given.
low-rank dimension D = 16
Dimension of the orthogonal projection space in preference editing (Equation 6). Selected from {4,8,16,32,64} based on validation performance (Section 4.4).
initial values of alpha and beta = (1, 0.5)
Initialization of the global adaptive fusion parameters in Equation (9). Chosen because it encodes the prior that positive relevance or preference usually leads to positive behavior (Section 4.4).
area classification threshold = top 20%
Threshold used to assign samples to Venn diagram areas in the visualization (Section 4.5). Not part of the core model but used to support the modeling-space claims.

assumptions (4)

domain assumption The causal graph includes an edge R->P (relevance influences preference).
Section 2.2 and Figure 1(a). The paper assumes this edge without testing it, and the entire preference editing component is designed to remove it.
domain assumption User behavior B is caused only by preference P and relevance R, with no unobserved confounders.
Section 2.2, Venn diagram. The six-area decomposition excludes B=1 when P=R=0 and B=0 when P=R=1, implying a deterministic causal structure.
domain assumption The relevance-to-preference intervention lies in a low-rank subspace of the last-layer representation and can be removed by an orthogonal projection O.
Section 3.2, Equation (6). Invoked from Geiger et al. [13], but not proven for these learned representations; no auxiliary loss enforces that O captures the true intervention.
domain assumption The global adaptive fusion parameters alpha and beta can separate the four relevance-preference statuses (P11, P10, P01, P00).
Section 3.3, Equation (9). The paper assumes the learned linear combination is sufficient to model the different areas.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Behavior Modeling Space Reconstruction for E-Commerce Search." pith.science (2026). https://pith.science/paper/S4LUPBU5

@misc{pith2026250118216,
  author       = {Pith},
  title        = {Pith review of: Behavior Modeling Space Reconstruction for E-Commerce Search},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/S4LUPBU5}},
  note         = {Machine review of arXiv:2501.18216}
}

read the original abstract

Delivering superior search services is crucial for enhancing customer experience and driving revenue growth. Conventionally, search systems model user behaviors by combining user preference and query item relevance statically, often through a fixed logical 'and' relationship. This paper reexamines existing approaches through a unified lens using both causal graphs and Venn diagrams, uncovering two prevalent yet significant issues: entangled preference and relevance effects, and a collapsed modeling space. To surmount these challenges, our research introduces a novel framework, DRP, which enhances search accuracy through two components to reconstruct the behavior modeling space. Specifically, we implement preference editing to proactively remove the relevance effect from preference predictions, yielding untainted user preferences. Additionally, we employ adaptive fusion, which dynamically adjusts fusion criteria to align with the varying patterns of relevance and preference, facilitating more nuanced and tailored behavior predictions within the reconstructed modeling space. Empirical validation on two public datasets and a proprietary search dataset underscores the superiority of our proposed methodology, demonstrating marked improvements in performance over existing approaches.

Figures

Figures reproduced from arXiv: 2501.18216 by the authors.

**Figure 2.** Framework Overview. The joint modeling framework is depicted in (a). The preference editing is represented in (b). [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 4.** Model visualization of adaptive fusion. the first to articulate the notion of relevance modeling, which has inspired numerous subsequent efforts aimed at designing pre-trained relevance models to be integrated within a preference modeling framework, operating as a joint behavior modeling approach. Techniques such as knowledge distillation [20], self-supervised learning [8], and pretraining pipelines [55] are empl… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

66 extracted references · 42 canonical work pages

[13]

Atticus Geiger, Zhengxuan Wu, Christopher Potts, Thomas Icard, and Noah Goodman. 2024. Finding alignments between interpretable causal variables and distributed neural representations. In Causal Learning and Reasoning . PMLR, 160–187

work page 2024
[1]

Qingyao Ai, Daniel N Hill, SVN Vishwanathan, and W Bruce Croft. 2019. A zero attention model for personalized product search. In Proceedings of the 28th ACM International Conference on Information and Knowledge Management . 379–388

2019
[2]

Qingyao Ai and Lakshmi Narayanan. R. 2021. Model-agnostic vs. model-intrinsic interpretability for explainable product search. In Proceedings of the 30th ACM International Conference on Information & Knowledge Management . 5–15

work page 2021
[3]

Qingyao Ai, Yongfeng Zhang, Keping Bi, Xu Chen, and W Bruce Croft. 2017. Learning a hierarchical embedding model for personalized product search. In Proceedings of the 40th International ACM SIGIR Conference on Research and Development in Information Retrieval . 645–654

2017
[4]

Qingyao Ai, Yongfeng Zhang, Keping Bi, and W Bruce Croft. 2019. Explainable product search with a dynamic relation embedding model. ACM Transactions on Information Systems (TOIS) 38, 1 (2019), 1–29

work page 2019
[5]

Keping Bi, Qingyao Ai, and W Bruce Croft. 2020. A transformer-based embedding model for personalized product search. In Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval . 1521–1524

2020
[6]

Keping Bi, Qingyao Ai, and W Bruce Croft. 2021. Learning a fine-grained review- based transformer model for personalized product search. In Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Infor- mation Retrieval. 123–132

work page 2021
[7]

David Carmel, Elad Haramaty, Arnon Lazerson, Liane Lewin-Eytan, and Yoelle Maarek. 2020. Why do people buy seemingly irrelevant items in voice product search? On the relation between product relevance and customer satisfaction in ecommerce. In Proceedings of the 13th international conference on web search and data mining. 79–87

work page 2020

Show all 66 references

[8]

Zeyuan Chen, Wei Chen, Jia Xu, Zhongyi Liu, and Wei Zhang. 2023. Beyond Se- mantics: Learning a Behavior Augmented Relevance Model with Self-supervised Learning. In Proceedings of the 32nd ACM International Conference on Information and Knowledge Management. 4516–4522

2023
[9]

Heng-Tze Cheng, Levent Koc, Jeremiah Harmsen, Tal Shaked, Tushar Chandra, Hrishi Aradhye, Glen Anderson, Greg Corrado, Wei Chai, Mustafa Ispir, et al
[10]

Shitong Dai, Jiongnan Liu, Zhicheng Dou, Haonan Wang, Lin Liu, Bo Long, and Ji-Rong Wen. 2023. Contrastive Learning for User Sequence Representation in Personalized Product Search. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining . 380–389

2023
[11]

Yingqiang Ge, Shuya Zhao, Honglu Zhou, Changhua Pei, Fei Sun, Wenwu Ou, and Yongfeng Zhang. 2020. Understanding echo chambers in e-commerce recom- mender systems. In Proceedings of the 43rd international ACM SIGIR conference on research and development in information retrieval...

2020
[12]

Atticus Geiger, Hanson Lu, Thomas Icard, and Christopher Potts. 2021. Causal abstractions of neural networks. Advances in Neural Information Processing Systems 34 (2021), 9574–9586

2021
[14]

Huifeng Guo, Ruiming Tang, Yunming Ye, Zhenguo Li, and Xiuqiang He. 2017. DeepFM: a factorization-machine based neural network for CTR prediction.arXiv preprint arXiv:1703.04247 (2017)

2017 arXiv
[15]

Ruocheng Guo, Lu Cheng, Jundong Li, P Richard Hahn, and Huan Liu. 2020. A survey of learning causality with data: Problems and methods. ACM Computing Surveys (CSUR) 53, 4 (2020), 1–37

2020
[16]

Zhaolin Hong, Haitao Wang, Chengjie Qian, Wei Chen, Tianqi He, Yajie Zou, Qiang Liu, and Xingxing Wang. 2024. PRINT: Personalized Relevance Incentive Network for CTR Prediction in Sponsored Search. In Companion Proceedings of the ACM on Web Conference 2024 . 190–195

2024
[17]

Jiahui Huang, Lan Zhang, Junhao Wang, Shanyang Jiang, Dongbo Huang, Cheng Ding, and Lan Xu. 2024. Utilizing Non-click Samples via Semi-supervised Learn- ing for Conversion Rate Prediction. In Proceedings of the 18th ACM Conference on Recommender Systems. 350–359

2024
[18]

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 Proceedings of the 22nd ACM international conference on Information & Knowledge Management . 2333–2338

2013
[19]

Vipin Jain, BINDOO Malviya, and SATYENDRA Arya. 2021. An overview of electronic commerce (e-Commerce). The journal of contemporary issues in business and government 27, 3 (2021), 665–670

2021
[20]

Yunjiang Jiang, Yue Shang, Ziyang Liu, Hongwei Shen, Yun Xiao, Sulong Xu, Wei Xiong, Weipeng Yan, and Di Jin. 2020. BERT2DNN: BERT distillation with massive unlabeled data for online e-commerce search. In 2020 IEEE International Conference on Data Mining (ICDM) . IEEE, 212–221

2020
[21]

Jae-woong Lee, Seongmin Park, Joonseok Lee, and Jongwuk Lee. 2022. Bilateral self-unbiased learning from biased implicit feedback. In Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval. 29–39

2022
[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. 101–111

2023
[23]

Xinhang Li, Zhaopeng Qiu, Xiangyu Zhao, Zihao Wang, Yong Zhang, Chunxiao Xing, and Xian Wu. 2022. Gromov-wasserstein guided representation learning for cross-domain recommendation. In Proceedings of the 31st ACM International Conference on Information & Knowledge Management . ...

2022
[24]

Xiaopeng Li, Fan Yan, Xiangyu Zhao, Yichao Wang, Bo Chen, Huifeng Guo, and Ruiming Tang. 2023. Hamur: Hyper adapter for multi-domain recommendation. In Proceedings of the 32nd ACM International Conference on Information and Knowledge Management. 1268–1277

2023
[25]

Xinhang Li, Xiangyu Zhao, Jiaxing Xu, Yong Zhang, and Chunxiao Xing. 2023. IMF: Interactive Multimodal Fusion Model for Link Prediction. In Proceedings of the Web Conference 2023

2023
[26]

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
[27]

Siyi Lin, Sheng Zhou, Jiawei Chen, Yan Feng, Qihao Shi, Chun Chen, Ying Li, and Can Wang. 2024. ReCRec: Reasoning the Causes of Implicit Feedback for Debiased Recommendation. ACM Transactions on Information Systems (2024)

2024
[28]

Weilin Lin, Xiangyu Zhao, Yejing Wang, Tong Xu, and Xian Wu. 2022. AdaFS: Adaptive Feature Selection in Deep Recommender System. In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining . 3309–3317

2022
[29]

Weilin Lin, Xiangyu Zhao, Yejing Wang, Yuanshao Zhu, and Wanyu Wang. 2023. AutoDenoise: Automatic Data Instance Denoising for Recommendations. In Proceedings of the ACM Web Conference 2023 . 1003–1011

2023
[30]

Haochen Liu, Da Tang, Ji Yang, Xiangyu Zhao, Hui Liu, Jiliang Tang, and Youlong Cheng. 2022. Rating Distribution Calibration for Selection Bias Mitigation in Recommendations. In Proceedings of the ACM Web Conference 2022 . 2048–2057

2022
[31]

Jiongnan Liu, Zhicheng Dou, Guoyu Tang, and Sulong Xu. 2023. Jdsearch: A personalized product search dataset with real queries and full interactions. In Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval . 2945–2952

2023
[32]

Qidong Liu, Jiaxi Hu, Yutian Xiao, Xiangyu Zhao, Jingtong Gao, Wanyu Wang, Qing Li, and Jiliang Tang. 2023. Multimodal recommender systems: A survey. Comput. Surveys (2023)

2023
[33]

Qidong Liu, Xian Wu, Xiangyu Zhao, Yejing Wang, Zijian Zhang, Feng Tian, and Yefeng Zheng. 2024. Large Language Models Enhanced Sequential Recommenda- tion for Long-tail User and Item. arXiv preprint arXiv:2405.20646 (2024)

2024 arXiv
[34]

Qidong Liu, Xian Wu, Xiangyu Zhao, Yuanshao Zhu, Derong Xu, Feng Tian, and Yefeng Zheng. 2024. When MOE Meets LLMs: Parameter Efficient Fine-tuning for Multi-task Medical Applications. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in...

2024
[35]

Qidong Liu, Xiangyu Zhao, Yuhao Wang, Yejing Wang, Zijian Zhang, Yuqi Sun, Xiang Li, Maolin Wang, Pengyue Jia, Chong Chen, et al. 2024. Large Language Model Enhanced Recommender Systems: Taxonomy, Trend, Application and Future. arXiv preprint arXiv:2412.13432 (2024)

2024 arXiv
[36]

Shuchang Liu, Qingpeng Cai, Bowen Sun, Yuhao Wang, Ji Jiang, Dong Zheng, Kun Gai, Peng Jiang, Xiangyu Zhao, and Yongfeng Zhang. 2023. Exploration and Regularization of the Latent Action Space in Recommendation. In Proceedings of the Web Conference 2023

2023
[37]

Ziru Liu, Jiejie Tian, Qingpeng Cai, Xiangyu Zhao, Jingtong Gao, Shuchang Liu, Dayou Chen, Tonghao He, Dong Zheng, Peng Jiang, et al. 2023. Multi-task recommendations with reinforcement learning. In Proceedings of the ACM Web Conference 2023. 1273–1282

2023
[38]

Thanh Nguyen, Nikhil Rao, and Karthik Subbian. 2020. Learning Robust Models for e-Commerce Product Search. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics . 6861–6869

2020
[39]

Stephen Robertson, Hugo Zaragoza, et al . 2009. The probabilistic relevance framework: BM25 and beyond. Foundations and Trends® in Information Retrieval 3, 4 (2009), 333–389

2009
[40]

Teng Shi, Zihua Si, Jun Xu, Xiao Zhang, Xiaoxue Zang, Kai Zheng, Dewei Leng, Yanan Niu, and Yang Song. 2024. UniSAR: Modeling User Transition Behaviors between Search and Recommendation. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development ...

2024
[41]

Zhongxiang Sun, Zihua Si, Xiaoxue Zang, Dewei Leng, Yanan Niu, Yang Song, Xiao Zhang, and Jun Xu. 2023. KuaiSAR: A Unified Search And Recommendation Dataset. In Proceedings of the 32nd ACM International Conference on Information and Knowledge Management. 5407–5411

2023
[42]

Hao Wang, Tai-Wei Chang, Tianqiao Liu, Jianmin Huang, Zhichao Chen, Chao Yu, Ruopeng Li, and Wei Chu. 2022. ESCM2: entire space counterfactual multi- task model for post-click conversion rate estimation. In Proceedings of the 45th WWW ’25, April 28-May 2, 2025, Sydney, NSW, Au...

2022
[43]

Ruoxi Wang, Bin Fu, Gang Fu, and Mingliang Wang. 2017. Deep & cross network for ad click predictions. In Proceedings of the ADKDD’17 . 1–7

2017
[44]

Yejing Wang, Zhaocheng Du, Xiangyu Zhao, Bo Chen, Huifeng Guo, Ruiming Tang, and Zhenhua Dong. 2023. Single-shot Feature Selection for Multi-task Recommendations. In Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval...

2023
[45]

Yejing Wang, Shen Ge, Xiangyu Zhao, Xian Wu, Tong Xu, Chen Ma, and Zhi Zheng. 2023. Doctor specific tag recommendation for online medical record management. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining . 5150–5161

2023
[46]

Yejing Wang, Dong Xu, Xiangyu Zhao, Zhiren Mao, Peng Xiang, Ling Yan, Yao Hu, Zijian Zhang, Xuetao Wei, and Qidong Liu. 2024. GPRec: Bi-level User Modeling for Deep Recommenders. arXiv preprint arXiv:2410.20730 (2024)

2024 arXiv
[47]

Yuhao Wang, Xiangyu Zhao, Bo Chen, Qidong Liu, Huifeng Guo, Huanshuo Liu, Yichao Wang, Rui Zhang, and Ruiming Tang. 2023. PLATE: A Prompt- Enhanced Paradigm for Multi-Scenario Recommendations. In Proceedings of the 46th International ACM SIGIR Conference on Research and Develo...

2023
[48]

Yejing Wang, Xiangyu Zhao, Tong Xu, and Xian Wu. 2022. Autofield: Automating feature selection in deep recommender systems. In Proceedings of the ACM Web Conference 2022. 1977–1986

2022
[49]

Zimu Wang, Yue He, Jiashuo Liu, Wenchao Zou, Philip S Yu, and Peng Cui
[50]

Zhiqiang Wang, Qingyun She, and Junlin Zhang. 2021. MaskNet: Introducing Feature-Wise Multiplication to CTR Ranking Models by Instance-Guided Mask. (2021)

2021
[51]

He sicheng Wang Yuxin, Sun Qingxuan. 2023. Moka Massive Mixed Embedding . https://huggingface.co/moka-ai/m3e-base

2023
[52]

Rong Xiao, Jianhui Ji, Baoliang Cui, Haihong Tang, Wenwu Ou, Yanghua Xiao, Jiwei Tan, and Xuan Ju. 2019. Weakly supervised co-training of query rewrit- ing and semantic matching for e-commerce. In Proceedings of the twelfth ACM international conference on web search and data m...

2019
[53]

Derong Xu, Wei Chen, Wenjun Peng, Chao Zhang, Tong Xu, Xiangyu Zhao, Xian Wu, Yefeng Zheng, Yang Wang, and Enhong Chen. 2024. Large language models for generative information extraction: A survey. Frontiers of Computer Science 18, 6 (2024), 186357

2024
[54]

Shaowei Yao, Jiwei Tan, Xi Chen, Keping Yang, Rong Xiao, Hongbo Deng, and Xiaojun Wan. 2021. Learning a product relevance model from click-through data in e-commerce. In Proceedings of the Web Conference 2021 . 2890–2899

2021
[55]

Wen Zan, Yaopeng Han, Xiaotian Jiang, Yao Xiao, Yang Yang, Dayao Chen, and Sheng Chen. 2023. SPM: Structured Pretraining and Matching Architectures for Relevance Modeling in Meituan Search. In Proceedings of the 32nd ACM International Conference on Information and Knowledge Ma...

2023
[56]

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. InProceedings of the 31st ACM International Conference on Information & Knowledge Management . 2508–2518

2022
[57]

Weinan Zhang, Tianming Du, and Jun Wang. 2016. Deep Learning over Multi- field Categorical Data: –A Case Study on User Response Prediction. In Advances in Information Retrieval: 38th European Conference on IR Research, ECIR 2016, Padua, Italy, March 20–23, 2016. Proceedings 38...

2016
[58]

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
[59]

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 Recommender Systems Conference . ACM, 95–103

2018
[60]

Xiangyu Zhao, Long Xia, Lixin Zou, Hui Liu, Dawei Yin, and Jiliang Tang. 2020. Whole-Chain Recommendations. In Proceedings of the 29th ACM International Conference on Information & Knowledge Management . 1883–1891

2020
[61]

Xiangyu Zhao, Liang Zhang, Zhuoye Ding, Long Xia, Jiliang Tang, and Dawei Yin
[62]

Zhi Zheng, Zhaopeng Qiu, Tong Xu, Xian Wu, Xiangyu Zhao, Enhong Chen, and Hui Xiong. 2022. CBR: Context Bias aware Recommendation for Debiasing User Modeling and Click Prediction. In Proceedings of the ACM Web Conference 2022 . 2268–2276

2022
[63]

Feng Zhu, Mingjie Zhong, Xinxing Yang, Longfei Li, Lu Yu, Tiehua Zhang, Jun Zhou, Chaochao Chen, Fei Wu, Guanfeng Liu, et al. 2023. DCMT: A Direct Entire- Space Causal Multi-Task Framework for Post-Click Conversion Estimation. In 2023 IEEE 39th International Conference on Data...

2023
[2016]

In Proceedings of the 1st workshop on deep learning for recommender systems

Wide & deep learning for recommender systems. In Proceedings of the 1st workshop on deep learning for recommender systems . 7–10
[2018]

In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining

Recommendations with Negative Feedback via Pairwise Deep Reinforce- ment Learning. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining . ACM, 1040–1048
[2022]

In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining

Invariant preference learning for general debiasing in recommendation. In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. 1969–1978

1969

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.

[1] [13]

Atticus Geiger, Zhengxuan Wu, Christopher Potts, Thomas Icard, and Noah Goodman. 2024. Finding alignments between interpretable causal variables and distributed neural representations. In Causal Learning and Reasoning . PMLR, 160–187

work page 2024

[2] [1]

Qingyao Ai, Daniel N Hill, SVN Vishwanathan, and W Bruce Croft. 2019. A zero attention model for personalized product search. In Proceedings of the 28th ACM International Conference on Information and Knowledge Management . 379–388

2019

[3] [2]

Qingyao Ai and Lakshmi Narayanan. R. 2021. Model-agnostic vs. model-intrinsic interpretability for explainable product search. In Proceedings of the 30th ACM International Conference on Information & Knowledge Management . 5–15

work page 2021

[4] [3]

Qingyao Ai, Yongfeng Zhang, Keping Bi, Xu Chen, and W Bruce Croft. 2017. Learning a hierarchical embedding model for personalized product search. In Proceedings of the 40th International ACM SIGIR Conference on Research and Development in Information Retrieval . 645–654

2017

[5] [4]

Qingyao Ai, Yongfeng Zhang, Keping Bi, and W Bruce Croft. 2019. Explainable product search with a dynamic relation embedding model. ACM Transactions on Information Systems (TOIS) 38, 1 (2019), 1–29

work page 2019

[6] [5]

Keping Bi, Qingyao Ai, and W Bruce Croft. 2020. A transformer-based embedding model for personalized product search. In Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval . 1521–1524

2020

[7] [6]

Keping Bi, Qingyao Ai, and W Bruce Croft. 2021. Learning a fine-grained review- based transformer model for personalized product search. In Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Infor- mation Retrieval. 123–132

work page 2021

[8] [7]

David Carmel, Elad Haramaty, Arnon Lazerson, Liane Lewin-Eytan, and Yoelle Maarek. 2020. Why do people buy seemingly irrelevant items in voice product search? On the relation between product relevance and customer satisfaction in ecommerce. In Proceedings of the 13th international conference on web search and data mining. 79–87

work page 2020