Pith. sign in

REVIEW 4 major objections 5 minor 50 references

GIST: Cross-Domain Click-Through Rate Prediction via Guided Content-Behavior Distillation

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read GIST claims that cross-domain ad CTR prediction improves when a joint content-behavior item representation, distilled from source-domain ranking attention, drives lifelong sequence search and similarity features, citing offline AUC gains…

desk verdict Plausible industrial mechanism with a coherent decoupled-training story; the evidence as reported is not yet enough for the SOTA claim. read the letter →

arxiv 2507.05142 v1 pith:T4N3GKY6 submitted 2025-07-07 cs.AI

Wei Xu , Haoran Li , Baoyuan Ou , Lai Xu , Yingjie Qin , Ruilong Su , Ruiwen Xu This is my paper
classification cs.AI
keywords cross-domainrecommendationCTRpredictionlifelongbehaviorsequencecontent-behaviorjointrepresentationcontrastivelearningmulti-modalonlineadvertisingdecoupledtraining
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

The paper sets out to show that a click-through rate (CTR) model for online advertising can be improved by borrowing knowledge from a user's long recommendation-domain history without jointly training the two domains. Its system, GIST, learns a joint item representation that combines content signals (item text and images) with behavior signals (how often and how users interact with items), guided by item pairs that the source recommendation model's fine-grained matching stage considers most relevant. This joint representation is used to retrieve a small set of highly relevant historical items from each user's lifelong sequence and to produce similarity scores and score distributions that are fed into the advertising model as features. The paper reports offline AUC gains over strong lifelong-sequence baselines and a 7-day online A/B test in which CTR rose 3.1406%, income rose 2.1505%, and cost per mille rose 1.2528%, which it interprets as evidence that the content-behavior joint representation transfers knowledge effectively.

What carries the argument

The load-bearing object is the joint representation produced by CBJT, formed as $u_i = \mathrm{MLP}_{cs}(\hat{ct}_i \otimes \hat{bh}_i)$, where $\hat{ct}_i$ and $\hat{bh}_i$ are gated content and behavior embeddings and $\otimes$ is the outer product. That representation carries the transfer: it is the query-side and item-side key for the General Search Unit's soft search over lifelong sequences, and its cosine similarities are the input to ASI. The second mechanism is the distillation signal: ESU attention scores select high-quality item-to-item pairs that train the fusion, so the source model's notion of relevance is distilled into the joint representation. The third is ASI, which replaces direct use of the joint embedding in the ad model with discretized similarity-score embeddings and an $M_2$-dimensional similarity distribution vector, keeping the target model's own embeddings independent.

What would settle it

Train the CBJT module with the same pipeline but with the ESU attention scores shuffled or replaced by random pair labels, then measure GSU Recall@K and end-to-end AUC; if the model still achieves the reported gains, the distillation signal is not what drives performance. Alternatively, evaluate retrieval against held-out ad clicks not derived from ESU attention, and check whether the reported Recall@K and AUC gains persist.

Watch

Extended reading notes

Core claim

The central claim is that content and behavior signals should be fused into a joint item representation, distilled with guidance from the source ranking model's Exact Search Unit (ESU) attention, and then used not as embedding vectors to fine-tune but as search keys and similarity features for the target-domain ad model. In the Content-Behavior Joint Training Module (CBJT), a content-based encoder (text and images) and a behavior-based encoder (aligned to sparse ID embeddings of high-interaction items) are trained with InfoNCE contrastive losses, then fused through a gated network and a cross network. The fused representation is supervised with high-confidence item pairs selected from the source ESU when its attention score exceeds a threshold (0.4 in production). The target advertising model decouples from the source by learning its own ID embeddings, while the General Search Unit uses the joint representation to retrieve the top 100 items from the lifelong recommendation sequence, and the Asymmetric Similarity Integration (ASI) strategy converts cosine similarities into discretized score embeddings and a similarity distribution vector for the target item. The paper's evidence that this design works is the offline AUC improvement and the online A/B test outcome.

Load-bearing premise

The load-bearing premise is that the source model's ESU attention scores are trustworthy item-to-item relevance labels, so pairs with attention above 0.4 teach the joint representation a correct notion of relevance; because the paper evaluates retrieval against those same ESU pairs, noisy or domain-mismatched labels would compromise the retrieval-quality numbers.

Editorial extensions

If this is right

  • Decoupled training means the source-domain recommendation model can keep ingesting new interaction data without forcing the target ad model to retrain or fine-tune.
  • Joint content-behavior representations provide balanced retrieval across popular and long-tail items; the paper reports larger gains for low-interaction items than high-interaction items.
  • Using similarity scores and similarity distributions as side features outperforms directly concatenating joint embeddings into the ID-based ad model.
  • The online A/B test reports a 3.1406% CTR increase, a 2.1505% income increase, and a 1.2528% CPM increase over the deployed baseline, with 0.4% income being the platform's stated significance bar.

Reading between the lines

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

  • Extending beyond the paper: a natural check is whether retrieval quality holds when the ground truth is actual ad clicks in the target domain rather than the same ESU pairs used to train CBJT, since the paper's Recall@K is measured against those pairs.
  • Extending beyond the paper: the same decoupled recipe could be applied to other sparse target surfaces, such as new ad formats, new item categories, or cold-start recommendation, where a rich source domain exists but target behavior is thin.
  • Extending beyond the paper: because the fixed threshold depends on the stability of ESU attention scores over time, a drift in the source ranking model would require re-selecting the threshold or re-distilling the joint representation.
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 / 5 minor

Summary. The paper proposes GIST, a cross-domain CTR prediction framework deployed at Xiaohongshu (RedNote). GIST decouples source-domain and target-domain training: a Content-Behavior Joint Training Module (CBJT) learns item representations by jointly aligning multi-modal content signals with behavioral ID embeddings, using item-to-item pairs distilled from the source-domain Exact Search Unit (ESU) as guidance; these representations are then used for lifelong sequence search in the target advertising domain, and an Asymmetric Similarity Integration (ASI) strategy injects similarity scores and their distribution into the target attention module. The authors report offline AUC gains over DIN and SIM baselines on a 1.6-billion-record dataset, retrieval Recall@K improvements in ablation studies, and a 7-day online A/B test with CTR +3.1406%, Income +2.1505%, and CPM +1.2528%.

Significance. If the claimed results are validated, the paper's central ideas are industrially meaningful: decoupling source and target training avoids the distribution mismatch of joint training and the update bottleneck of pre-training/fine-tuning, and the CBJT/ASI design offers a practical way to combine content and behavior signals for lifelong sequence modeling. The paper's strength is that it addresses a real deployment setting with web-scale data and reports an online experiment, which is rare in academic papers. However, the evidence as reported is not yet sufficient to support the headline claims. The offline AUC differences are very small and are presented without variance or significance tests; the online A/B test lacks standard statistical reporting; and the retrieval evaluation used to justify several module design choices is circular. The baseline set is narrow, and the central decoupling claim is not tested against joint training or pre-training/fine-tuning paradigms. These issues are load-bearing for the paper's claim of state-of-the-art performance, so the manuscript needs substantive revision rather than minor polishing.

major comments (4)
  1. [§3.1.3 and §5.1.3] The retrieval evaluation is circular. In §3.1.3, the Union Training Model is trained using ESU item-to-item pairs as positive guidance, where the pairs are selected as (target item, most relevant historical item) with attention score above threshold θ. In §5.1.3, the same ESU pairs are used as ground truth for Recall@K. Consequently, a model trained to make these pairs similar will trivially retrieve them well, and the Recall numbers in Tables 3 and 5 and Figure 6 largely measure how faithfully the model memorized its own training signal rather than measuring general retrieval quality. This undermines the module-level conclusions drawn from Recall@K, including the choice of θ. Please re-evaluate retrieval against a held-out or independently constructed ground truth—for example, ESU pairs from a time period after the CBJT training data, or manually judged relevance pairs—and present the updated results.
  2. [§5.2, Table 1, and §5.5] The central performance claims are reported without statistical support. Table 1 lists AUC and relative AUC gains (e.g., GIST 0.7720 vs. SIM Soft attention 0.7701) but gives no standard deviations, confidence intervals, seeds, or significance tests; on a 1.6-billion-sample dataset a 0.0019 AUC gap is likely statistically significant, but the reader cannot verify this, and the gap is small enough that run-to-run or day-to-day variation matters. Similarly, the online A/B test in §5.5 reports 3.1406% CTR, 2.1505% Income, and 1.2528% CPM improvements but does not state the control model, traffic split, user sample sizes, confidence intervals, or p-values. The sentence 'even a 0.4% improvement in income is considered statistically significant' asserts an internal convention without demonstrating that the reported 2.15% is significant under that convention. Please provide uncertainty quantification for all headline numbers, including bootstrapped confidence intervals for the offline AUC and details of the online test design.
  3. [§5.1.2] The baseline set is too narrow to support the claim of surpassing state-of-the-art methods. Only DIN and three SIM variants are compared; the related-work section itself cites recent cross-domain lifelong sequence models, notably Hou et al. [12] (Cross-Domain Lifelong Sequential Modeling, KDD 2024), as well as other industrial sequence models such as TWIN [3], and these are not included in the offline comparison. Since the paper's central claim is 'surpassing SOTA methods', the evaluation must include at least the closest recent cross-domain lifelong-sequence baselines, ideally also a joint-training cross-domain CDR baseline and a pre-training/fine-tuning baseline. Without these, the improvement over SIM variants may reflect an incomplete comparison rather than a true advance.
  4. [§3 and §5.3] The decoupled training paradigm, which is a headline contribution, is never directly tested. The introduction motivates GIST by arguing that joint training and pre-training/fine-tuning underperform in the industrial setting, but all ablations in §5.3 vary encoders, fusion components, learning signals, or integration strategies within the GIST framework; there is no experiment that instantiates a joint-training variant or a pre-training-then-fine-tuning variant of the same architecture and reports its performance against GIST. The claimed advantage of decoupling is therefore supported only by argument. Please add an experiment that compares GIST with an equivalent model trained in a joint or pre-train/fine-tune manner, or explicitly state that such a comparison is outside the scope and temper the claim accordingly.
minor comments (5)
  1. [§5.1.5] Several hyperparameters that affect the method are not specified, including the interval counts M1 and M2 for the ASI similarity discretization, the exact interaction-count threshold for the behavior-based encoder, the embedding dimension for the joint representation, and the number of training steps or learning rates. Providing these values would improve reproducibility.
  2. [§5.3.1] There is a typo in the first sentence of the 'Analyses of the CBJT Module' subsection: 'we design ablation experiments to to demonstrate' should read 'to demonstrate'.
  3. [Figure 2 and Figure 3] The two figures are dense and the labels (e.g., 'Gate NN', 'CrossNetwork', 'Union Training Model', 'Distribution Emb') are difficult to read at the current resolution. Please enlarge the font and consider separating the CBJT and ASI schematic into sub-figures with clearer arrows.
  4. [§2, Equation (2)] The notation S_A and H_B is introduced in the problem definition but the roles of these two short behavior sequences (source-domain and target-domain) are not explained until the methodology section; a brief clarification at the point of definition would improve readability.
  5. [References and formatting] The paper still contains ACM template placeholder text ('Make sure to enter the correct conference title from your rights confirmation email') and an outdated 2018 copyright line in the reference format; these should be cleaned up before submission.

Circularity Check

1 steps flagged · score 6.0 of 10

Retrieval Recall@K is evaluated against the same ESU i2i pairs used to train CBJT, making the retrieval evidence circular; AUC results remain independent.

  1. fitted input called prediction [Section 3.1.3 (Union Training Model) and Section 5.1.3 (Metrics), used in Tables 3/5 and Figure 6]
    "we choose to distill ESU pairs consisting of the target and the most relevant item to enhance the fusion process. ... Finally, we align the previously distilled ESU pairs and optimize the model using an in-batch contrastive loss function. ... To assess the retrieval performance of the joint representations, we index the joint representations of all items, using ESU pairs as the ground truth. We evaluate the performance using Recall@K at various K levels. ..."

    CBJT is trained with an in-batch contrastive loss whose positives are exactly the ESU i2i pairs (target item, highest-attention historical item). The retrieval metric in Section 5.1.3 then uses those same ESU pairs as ground truth for Recall@K. High Recall@K therefore measures how well the learned representation reproduces its own training labels; it is a fit-accuracy report, not an independent measure of retrieval quality. Since the paper says it switches to Recall for subsequent experiments, this circular metric is load-bearing for the CBJT ablations (Tables 3 and 5) and for the ESU-pair threshold selection in Figure 6. The AUC comparisons on click labels are independent and still support GIST.

full rationale

The main CTR claim is supported by AUC on 1.6B click records, which is independent of the ESU-pair supervision, so the central result is not definitionally forced. However, the paper presents Recall@K as evidence that the CBJT joint representation has strong retrieval ability, and that claim is circular: the representation is optimized to rank ESU pairs high, and Recall@K is computed against those same ESU pairs as ground truth. Moreover, the paper explicitly adopts Recall as the metric for subsequent experiments, making the circular metric load-bearing for the ablation of the union training module and the choice of the ESU-pair threshold. The A/B test lacks confidence intervals and control-arm detail, but that is a statistical-evidence weakness rather than a circularity. Self-citations in the paper (e.g., [16]) are not load-bearing for the method. Overall: one concrete reduction-by-construction in the retrieval evaluation, while the AUC-based offline comparison and online revenue claim retain independent content; hence a partial-circularity score of 6.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new physical or conceptual entities. All assumptions are domain assumptions about data and transferability, plus a design choice about decoupled training. The free parameters are data-selection thresholds and the top-K size, tuned on proprietary industrial data. The undefined M1/M2 interval counts weaken the reproducibility of ASI.

free parameters (4)
  • ESU pair selection threshold theta = 0.4
    The threshold for filtering ESU i2i pairs is chosen by ablation on the i2i dataset (Figure 6), testing 0.2, 0.3, 0.4, 0.5 and picking 0.4. This is a data-fitted hyperparameter.
  • Interaction count threshold for behavior alignment = 200,000
    Sparse ID embeddings with interaction counts greater than 200,000 are used as alignment targets in the behavior-based encoder (Section 5.1.5). The value is chosen by the authors and not justified independently.
  • Top-K for cross-domain GSU search = 100
    The number of lifelong behavior sequences retrieved for each target item is set to 100 (Section 5.1.5). No sensitivity analysis is shown for this value.
  • Number of intervals M1 and M2 = Not specified
    The ASI module partitions the similarity range into M1 intervals for score embedding and M2 intervals for the distribution vector, but the actual values are never given in the paper. This is an undefined free parameter.
assumptions (5)
  • domain assumption ESU attention scores from the source recommendation model are reliable item-to-item relevance labels.
    The CBJT module is trained to distill ESU pairs (target item and highest-attention historical item). The paper assumes these pairs are high-quality signal (Section 3.1.3).
  • domain assumption Recommendation-domain behavior sequences transfer to advertising-domain CTR prediction.
    The core premise of the paper is that users' long recommendation histories are informative for ad clicks despite different user intent (Section 1, Figure 1).
  • ad hoc to paper Decoupled training with frozen joint representations outperforms joint training or pre-training with fine-tuning.
    The paper argues decoupling is more stable (Section 3.2) but does not empirically compare against joint training or fine-tuning in the experiments. This is a design assumption.
  • domain assumption Source-domain item ID embeddings contain behavioral information that can be distilled into content embeddings.
    The behavior-based encoder aligns content embeddings with sparse ID embeddings for high-interaction items, assuming the ID embeddings capture collaborative signals worth transferring (Section 3.1.2).
  • domain assumption The online A/B test is a valid measure of causal CTR improvement.
    The paper reports large online gains without stating randomization, sample sizes, or statistical significance methodology (Section 5.5).

how reviews work

0 comments
Cite this review

Pith. "Pith review of GIST: Cross-Domain Click-Through Rate Prediction via Guided Content-Behavior Distillation." pith.science (2026). https://pith.science/paper/T4N3GKY6

@misc{pith2026250705142,
  author       = {Pith},
  title        = {Pith review of: GIST: Cross-Domain Click-Through Rate Prediction via Guided Content-Behavior Distillation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/T4N3GKY6}},
  note         = {Machine review of arXiv:2507.05142}
}
read the original abstract

Cross-domain Click-Through Rate prediction aims to tackle the data sparsity and the cold start problems in online advertising systems by transferring knowledge from source domains to a target domain. Most existing methods rely on overlapping users to facilitate this transfer, often focusing on joint training or pre-training with fine-tuning approach to connect the source and target domains. However, in real-world industrial settings, joint training struggles to learn optimal representations with different distributions, and pre-training with fine-tuning is not well-suited for continuously integrating new data. To address these issues, we propose GIST, a cross-domain lifelong sequence model that decouples the training processes of the source and target domains. Unlike previous methods that search lifelong sequences in the source domains using only content or behavior signals or their simple combinations, we innovatively introduce a Content-Behavior Joint Training Module (CBJT), which aligns content-behavior distributions and combines them with guided information to facilitate a more stable representation. Furthermore, we develop an Asymmetric Similarity Integration strategy (ASI) to augment knowledge transfer through similarity computation. Extensive experiments demonstrate the effectiveness of GIST, surpassing SOTA methods on offline evaluations and an online A/B test. Deployed on the Xiaohongshu (RedNote) platform, GIST effectively enhances online ads system performance at scale, serving hundreds of millions of daily active users.

Figures

Figures reproduced from arXiv: 2507.05142 by the authors.

Figure 1
Figure 1. Illustration of cross-domain recommendation. Here, [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Overview of GIST, a Guided Content-Behavior Distillation framework designed to facilitate knowledge transfer in [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 2
Figure 2. This module combines behavioral signals and content [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figures from the paper (4 more)
Figure 3
Figure 3. Figure 3: Structure of the Asymmetric Similarity Integration [PITH_FULL_IMAGE:figures/full_fig_p005_3.png]
Figure 4
Figure 4. Figure 4: An overview of GIST workflow in the online system. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: The variation of attention distribution over time. [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: We choose ESU pairs based on different threshold [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

50 extracted references · 21 canonical work pages

  1. [12]

    Ruijie Hou, Zhaoyang Yang, Yu Ming, Hongyu Lu, Zhuobin Zheng, Yu Chen, Qinsong Zeng, and Ming Chen. 2024. Cross-Domain LifeLong Sequential Model- ing for Online Click-Through Rate Prediction. In Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining . 5116–5125

    work page 2024

  2. [3]

    Jianxin Chang, Chenbin Zhang, Zhiyi Fu, Xiaoxue Zang, Lin Guan, Jing Lu, Yiqun Hui, Dewei Leng, Yanan Niu, Yang Song, et al. 2023. TWIN: TWo-stage interest network for lifelong user behavior modeling in CTR prediction at kuaishou. In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. 3785–3794

    2023

  3. [1]

    Jiangxia Cao, Xixun Lin, Xin Cong, Jing Ya, Tingwen Liu, and Bin Wang. 2022. Disencdr: Learning disentangled representations for cross-domain recommenda- tion. In Proceedings of the 45th International ACM SIGIR conference on research and development in information retrieval . 267–277

    work page 2022

  4. [2]

    Yue Cao, Xiaojiang Zhou, Jiaqi Feng, Peihao Huang, Yao Xiao, Dayao Chen, and Sheng Chen. 2022. Sampling is all you need on modeling long-term user behaviors for CTR prediction. In Proceedings of the 31st ACM International Conference on Information & Knowledge Management . 2974–2983

    2022

  5. [4]

    Lei Chen, Fajie Yuan, Jiaxi Yang, Xiangnan He, Chengming Li, and Min Yang

    work page

  6. [5]

    Qiwei Chen, Changhua Pei, Shanshan Lv, Chao Li, Junfeng Ge, and Wenwu Ou

  7. [6]

    Shu Chen, Zitao Xu, Weike Pan, Qiang Yang, and Zhong Ming. 2024. A Survey on Cross-Domain Sequential Recommendation. arXiv preprint arXiv:2401.04971 (2024)

    arXiv 2024

  8. [7]

    arXiv preprint arXiv:2108.04468 (2021)

    End-to-end user behavior retrieval in click-through rateprediction model. arXiv preprint arXiv:2108.04468 (2021)

    arXiv 2021

Show all 50 references
  1. [8]

    Xinlei Chen and Kaiming He. 2021. Exploring simple siamese representation learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 15750–15758

    2021

  2. [9]

    Ting Chen, Simon Kornblith, Mohammad Norouzi, and Geoffrey Hinton. 2020. A simple framework for contrastive learning of visual representations. In Interna- tional conference on machine learning . PMLR, 1597–1607

    2020

  3. [10]

    Ningya Feng, Junwei Pan, Jialong Wu, Baixu Chen, Ximei Wang, Qian Li, Xian Hu, Jie Jiang, and Mingsheng Long. 2024. Long-Sequence Recommendation Models Need Decoupled Embeddings. arXiv preprint arXiv:2410.02604 (2024)

    2024 arXiv

  4. [11]

    Xiuqi Deng, Lu Xu, Xiyao Li, Jinkai Yu, Erpeng Xue, Zhongyuan Wang, Di Zhang, Zhaojie Liu, Guorui Zhou, Yang Song, et al. 2024. End-to-end training of Multimodal Model and ranking Model. arXiv preprint arXiv:2404.06078 (2024)

    2024 arXiv

  5. [13]

    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

  6. [14]

    Dongfang Li, Baotian Hu, Qingcai Chen, Xiao Wang, Quanchang Qi, Liubin Wang, and Haishan Liu. 2021. Attentive capsule network for click-through rate and conversion rate prediction in online advertising. Knowledge-based systems 211 (2021), 106522

    2021

  7. [15]

    Guangneng Hu, Yu Zhang, and Qiang Yang. 2018. Conet: Collaborative cross networks for cross-domain recommendation. In Proceedings of the 27th ACM international conference on information and knowledge management . 667–676

    2018

  8. [16]

    Haoran Li, Yingjie Qin, Baoyuan Ou, Lai Xu, and Ruiwen Xu. 2025. HoPE: Hybrid of Position Embedding for Length Generalization in Vision-Language Models. arXiv preprint arXiv:2505.20444 (2025)

    2025

  9. [17]

    Haoran Li, Junqi Liu, Zexian Wang, Shiyuan Luo, Xiaowei Jia, and Huaxiu Yao

  10. [18]

    Junnan Li, Dongxu Li, Caiming Xiong, and Steven Hoi. 2022. Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation. In International conference on machine learning . PMLR, 12888–12900

    2022

  11. [19]

    Pan Li and Alexander Tuzhilin. 2020. Ddtcdr: Deep dual transfer cross domain recommendation. In Proceedings of the 13th International Conference on Web Search and Data Mining . 331–339

    2020

  12. [20]

    Junnan Li, Dongxu Li, Silvio Savarese, and Steven Hoi. 2023. Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models. In International conference on machine learning . PMLR, 19730–19742

    2023

  13. [21]

    Hao Liu, Lei Guo, Lei Zhu, Yongqiang Jiang, Min Gao, and Hongzhi Yin. 2024. MCRPL: A Pretrain, Prompt, and Fine-tune Paradigm for Non-overlapping Many- to-one Cross-domain Recommendation.ACM Transactions on Information Systems 42, 4 (2024), 1–24

    2024

  14. [22]

    Meng Liu, Jianjun Li, Guohui Li, and Peng Pan. 2020. Cross domain recom- mendation via bi-directional transfer graph collaborative filtering networks. In Proceedings of the 29th ACM international conference on information & knowledge management. 885–894

    2020

  15. [23]

    Yakun Li, Lei Hou, and Juanzi Li. 2023. Preference-aware graph attention net- works for cross-domain recommendations with collaborative knowledge graph. ACM Transactions on Information Systems 41, 3 (2023), 1–26

    2023

  16. [24]

    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 Proceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining . 1930–1939

    2018

  17. [25]

    2019.𝜋-net: A parallel information-sharing network for shared-account cross- domain sequential recommendations

    Muyang Ma, Pengjie Ren, Yujie Lin, Zhumin Chen, Jun Ma, and Maarten de Rijke. 2019.𝜋-net: A parallel information-sharing network for shared-account cross- domain sequential recommendations. In Proceedings of the 42nd international ACM SIGIR conference on research and developme...

    2019

  18. [26]

    Qi Liu, Xuyang Hou, Defu Lian, Zhe Wang, Haoran Jin, Jia Cheng, and Jun Lei. 2024. AT4CTR: Auxiliary Match Tasks for Enhancing Click-Through Rate Prediction. In Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 38. 8787–8795

    2024

  19. [27]

    Wentao Ouyang, Xiuwu Zhang, Shukui Ren, Li Li, Kun Zhang, Jinmei Luo, Zhaojie Liu, and Yanlong Du. 2021. Learning graph meta embeddings for cold-start ads in click-through rate prediction. In Proceedings of the 44th International ACM SIGIR Conference on Research and Developmen...

    2021

  20. [28]

    Wentao Ouyang, Xiuwu Zhang, Lei Zhao, Jinmei Luo, Yu Zhang, Heng Zou, Zhaojie Liu, and Yanlong Du. 2020. Minet: Mixed interest network for cross- domain click-through rate prediction. InProceedings of the 29th ACM international conference on information & knowledge management ...

    2020

  21. [29]

    Aaron van den Oord, Yazhe Li, and Oriol Vinyals. 2018. Representation learning with contrastive predictive coding. arXiv preprint arXiv:1807.03748 (2018)

    2018 arXiv

  22. [30]

    Jiarui Qin, Weinan Zhang, Xin Wu, Jiarui Jin, Yuchen Fang, and Yong Yu. 2020. User behavior retrieval for click-through rate prediction. InProceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval. 2347–2356

    2020

  23. [31]

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. 2021. Learning transferable visual models from natural language supervision. In International conference on machine learni...

    2021

  24. [32]

    Qi Pi, Guorui Zhou, Yujing Zhang, Zhe Wang, Lejian Ren, Ying Fan, Xiaoqiang Zhu, and Kun Gai. 2020. Search-based user interest modeling with lifelong sequential behavior data for click-through rate prediction. In Proceedings of the 29th ACM International Conference on Informat...

    2020

  25. [33]

    Wenchao Sun, Muyang Ma, Pengjie Ren, Yujie Lin, Zhumin Chen, Zhaochun Ren, Jun Ma, and Maarten De Rijke. 2021. Parallel split-join networks for shared account cross-domain sequential recommendations. IEEE Transactions on Knowl- edge and Data Engineering 35, 4 (2021), 4106–4123...

    2021

  26. [34]

    Pawel Swietojanski, Jinyu Li, and Steve Renals. 2016. Learning hidden unit con- tributions for unsupervised acoustic model adaptation. IEEE/ACM Transactions on Audio, Speech, and Language Processing 24, 8 (2016), 1450–1463

    2016

  27. [35]

    Xiang-Rong Sheng, Feifan Yang, Litong Gong, Biao Wang, Zhangming Chan, Yujing Zhang, Yueyao Cheng, Yong-Nan Zhu, Tiezheng Ge, Han Zhu, et al

  28. [36]

    InProceedings of the 33rd ACM International Conference on Information and Knowledge Management

    Enhancing Taobao Display Advertising with Multimodal Representations: Challenges, Approaches and Insights. InProceedings of the 33rd ACM International Conference on Information and Knowledge Management . 4858–4865

  29. [37]

    Jianwei Yang, Yonatan Bisk, and Jianfeng Gao. 2021. Taco: Token-aware cascade contrastive learning for video-text alignment. In Proceedings of the IEEE/CVF international conference on computer vision . 11562–11572

    2021

  30. [38]

    Jinyu Yang, Jiali Duan, Son Tran, Yi Xu, Sampath Chanda, Liqun Chen, Belinda Zeng, Trishul Chilimbi, and Junzhou Huang. 2022. Vision-language pre-training with triple contrastive learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . 1...

    2022

  31. [39]

    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

  32. [40]

    Ruoxi Wang, Rakesh Shivanna, Derek Cheng, Sagar Jain, Dong Lin, Lichan Hong, and Ed Chi. 2021. Dcn v2: Improved deep & cross network and practical lessons for web-scale learning to rank systems. In Proceedings of the web conference 2021 . 1785–1797

    2021

  33. [41]

    Amir Zadeh, Minghai Chen, Soujanya Poria, Erik Cambria, and Louis-Philippe Morency. 2017. Tensor fusion network for multimodal sentiment analysis. arXiv preprint arXiv:1707.07250 (2017)

    2017 arXiv

  34. [42]

    Shuai Zhang, Lina Yao, Aixin Sun, and Yi Tay. 2019. Deep learning based recom- mender system: A survey and new perspectives. ACM computing surveys (CSUR) 52, 1 (2019), 1–38

    2019

  35. [43]

    Tianhe Yu, Saurabh Kumar, Abhishek Gupta, Sergey Levine, Karol Hausman, and Chelsea Finn. 2020. Gradient surgery for multi-task learning. Advances in Neural Information Processing Systems 33 (2020), 5824–5836

    2020

  36. [44]

    Zheng Yuan, Fajie Yuan, Yu Song, Youhua Li, Junchen Fu, Fei Yang, Yunzhu Pan, and Yongxin Ni. 2023. Where to go next for recommender systems? id- vs. modality-based recommender models revisited. In Proceedings of the 46th International ACM SIGIR Conference on Research and Deve...

    2023

  37. [45]

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

    2018

  38. [46]

    Yongchun Zhu, Zhenwei Tang, Yudan Liu, Fuzhen Zhuang, Ruobing Xie, Xu Zhang, Leyu Lin, and Qing He. 2022. Personalized transfer of user preferences for cross-domain recommendation. In Proceedings of the fifteenth ACM international conference on web search and data mining . 1507–1515

    2022

  39. [47]

    Yujing Zhang, Zhangming Chan, Shuhao Xu, Weijie Bian, Shuguang Han, Hongbo Deng, and Bo Zheng. 2022. KEEP: An industrial pre-training framework for online recommendation via knowledge extraction and plugging. In Proceedings of the 31st ACM International Conference on Informati...

    2022

  40. [48]

    Chuang Zhao, Hongke Zhao, Ming He, Jian Zhang, and Jianping Fan. 2023. Cross- domain recommendation via user interest alignment. In Proceedings of the ACM Web Conference 2023. 887–896

    2023

  41. [2021]

    IEEE Transactions on Knowledge and Data Engineering 35, 3 (2021), 3239–3252

    User-specific adaptive fine-tuning for cross-domain recommendations. IEEE Transactions on Knowledge and Data Engineering 35, 3 (2021), 3239–3252

    2021

  42. [2024]

    arXiv preprint arXiv:2404.01165 (2024)

    LITE: Modeling Environmental Ecosystems with Multimodal Large Lan- guage Models. arXiv preprint arXiv:2404.01165 (2024)

    2024 arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.