pith. sign in

arxiv: 2511.19931 · v2 · pith:CV2VEDJPnew · submitted 2025-11-25 · 💻 cs.IR · cs.AI

LLM-EDT: Large Language Model Enhanced Cross-domain Sequential Recommendation with Dual-phase Training

Ziwei Liu , Qidong Liu , Wanyu Wang , Yejing Wang , Pengyue Jia , Tong Xu , Wei Huang , Chong Chen
show 1 more author
Xiangyu Zhao
This is my paper

Pith reviewed 2026-05-21 18:51 UTC · model grok-4.3

classification 💻 cs.IR cs.AI
keywords cross-domain sequential recommendationlarge language modelsdual-phase trainingitem augmentationuser profilingrecommendation systemstransfer learning
0
0 comments X

The pith

LLM-EDT improves cross-domain sequential recommendation by using a transferable item augmenter, dual-phase training, and domain-aware profiling to fix imbalance, transition, and profiling problems.

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

This paper proposes LLM-EDT to enhance cross-domain sequential recommendation with large language models. It targets the imbalance where one domain's interactions dominate and obscure features from others, the transition problem of capturing preferences in mixed sequences, and rough user profiling that lacks domain detail. The solution generates adaptive cross-domain behaviors to balance data with limited noise, applies dual-phase training to combine shared background knowledge with domain-specific threads, and builds profiles by summarizing each domain then aggregating them. A reader would care because these fixes could produce more accurate next-item suggestions when users move between domains such as movies and products. Tests on three public datasets support the gains over prior approaches.

Core claim

The paper claims that large language models, acting as generators and encoders, can be combined with a transferable item augmenter that creates relevant cross-domain behaviors, a dual-phase training strategy that first builds a domain-shared background then strengthens domain-specific threads, and a domain-aware profiling module that summarizes preferences per domain before adaptive aggregation, thereby reducing dominance by any single domain, easing capture of cross-domain transitions in sequences, and yielding richer user profiles for improved next-item prediction in each domain.

What carries the argument

The transferable item augmenter that adaptively generates possible cross-domain behaviors using the LLM to address imbalance with reduced noise, the dual-phase training strategy that empowers domain-specific modeling with shared context, and the domain-aware profiling module that summarizes and aggregates per-domain preferences.

If this is right

  • The imbalance issue is mitigated because generated cross-domain items supplement the weaker domain without adding much irrelevant noise.
  • The transition issue is reduced as dual-phase training lets the model learn shared patterns first then refine domain-specific predictions in mixed sequences.
  • User profiles become more comprehensive through per-domain summarization followed by adaptive aggregation, supporting better personalization.
  • Next-item prediction accuracy rises in each domain on the tested datasets compared with earlier cross-domain methods.
  • The overall framework demonstrates effectiveness when evaluated on three standard public datasets for cross-domain sequential recommendation.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The augmentation approach might extend to other settings where one data source dominates, such as multi-platform user modeling.
  • Testing the same modules with smaller or distilled language models could check if the gains hold at lower cost.
  • The dual-phase structure suggests a general pattern for combining broad knowledge with narrow task focus in sequential models.

Load-bearing premise

The large language model can generate cross-domain user behaviors that are relevant and low in noise enough to boost recommendation performance without extra human validation or filtering of the outputs.

What would settle it

An experiment on one of the public datasets that shows next-item prediction accuracy in the target domain stays the same or falls after adding the LLM-generated items from the transferable augmenter would indicate the augmentation step does not deliver the claimed benefit.

Figures

Figures reproduced from arXiv: 2511.19931 by Chong Chen, Pengyue Jia, Qidong Liu, Tong Xu, Wanyu Wang, Wei Huang, Xiangyu Zhao, Yejing Wang, Ziwei Liu.

Figure 1
Figure 1. Figure 1: Illustration for imbalanced domain distribution. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Framework of proposed LLM-EDT where E𝑖 , E𝑠 𝐴 𝑘 and E𝑠 𝐵 𝑘 respectively represent the embeddings of item 𝑖, 𝑠 𝐴 𝑘 and 𝑠 𝐵 𝑘 . 𝑠 𝐴 𝑖 , 𝑠𝐵 𝑖 respectively represent the corresponding position for 𝑠𝑐𝑜𝑟𝑒𝐴 𝑖 , 𝑠𝑐𝑜𝑟𝑒𝐵 𝑖 . For the A2B perspective, as discussed previously, our goal is to generate the user’s latent interest in do￾main B by feeding the representative domain A interactions into LLMs. Thus, for a gener… view at source ↗
Figure 3
Figure 3. Figure 3: Distribution Comparison in Cloth - Sport. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Performance on grouped users in Cloth-Sports [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Hyper-parameter Results on Amazon Cloth-Sport. [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Case Study on User 7382 in Cloth-Sport. B.3 Generality Validation To validate the generality of our proposed framework, we compare the most competitive baseline, i.e., LLM4CDSR, with our proposed LLM-EDT by replacing the backbone with two traditional models, i.e., GRU4Rec and BERT4Rec, shown in [PITH_FULL_IMAGE:figures/full_fig_p013_6.png] view at source ↗
read the original abstract

Cross-domain Sequential Recommendation (CDSR) has been proposed to enrich user-item interactions by incorporating information from various domains. Despite current progress, the imbalance issue and transition issue hinder further development of CDSR. The former one presents a phenomenon that the interactions in one domain dominate the entire behavior, leading to difficulty in capturing the domain-specific features in the other domain. The latter points to the difficulty in capturing users' cross-domain preferences within the mixed interaction sequence, resulting in poor next-item prediction performance for specific domains. With world knowledge and powerful reasoning ability, Large Language Models (LLMs) partially alleviate the above issues by performing as a generator and an encoder. However, current LLMs-enhanced CDSR methods are still under exploration, which fail to recognize the irrelevant noise and rough profiling problems. Thus, to make peace with the aforementioned challenges, we proposed an LLMs Enhanced Cross-domain Sequential Recommendation with Dual-phase Training ({LLM-EDT}). To address the imbalance issue while introducing less irrelevant noise, we first propose the transferable item augmenter to adaptively generate possible cross-domain behaviors for users. Then, to alleviate the transition issue, we introduce a dual-phase training strategy to empower the domain-specific thread with a domain-shared background. As for the rough profiling problem, we devise a domain-aware profiling module to summarize the user's preference in each domain and adaptively aggregate them to generate comprehensive user profiles. The experiments on three public datasets validate the effectiveness of our proposed LLM-EDT. To ease reproducibility, we have released the detailed code online at {https://anonymous.4open.science/r/LLM-EDT-583F}.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper proposes LLM-EDT for cross-domain sequential recommendation (CDSR). It uses LLMs to mitigate the imbalance issue (via a transferable item augmenter that generates cross-domain user behaviors) and the transition issue (via dual-phase training that combines domain-specific and shared modeling). A domain-aware profiling module is added to improve user preference summarization across domains. Effectiveness is claimed based on experiments on three public datasets, with code released for reproducibility.

Significance. If the reported gains prove robust and attributable to the proposed modules rather than unvalidated LLM outputs or other factors, the work could meaningfully advance LLM-augmented CDSR by providing targeted solutions to data imbalance and cross-domain preference modeling. The explicit code release is a positive contribution to reproducibility in this area.

major comments (2)
  1. [Transferable item augmenter description] The central claim for the transferable item augmenter (described in the method section following the abstract) is that it addresses imbalance while introducing less irrelevant noise. This rests on the unverified assumption that LLM-generated cross-domain behaviors are domain-plausible and low-noise. No human evaluation, catalog-overlap metric, noise quantification, or ablation isolating the augmenter from dual-phase training and profiling is reported, making it impossible to confirm the noise-reduction benefit or rule out that gains derive from simply adding more data.
  2. [Experiments and results] The experimental validation (results section) reports effectiveness on three datasets but provides no error bars, statistical significance tests, or full ablation tables isolating each module. Without these, the cross-domain performance claims cannot be assessed for robustness against the reader's noted concerns about post-hoc choices or data leakage.
minor comments (2)
  1. [Abstract] The anonymous code link in the abstract is standard for blind review but should be replaced with a permanent repository (e.g., GitHub) in the camera-ready version to support the reproducibility claim.
  2. [Dual-phase training strategy] Notation for the dual-phase training (e.g., how domain-specific threads and shared background are combined in the loss) could be clarified with a single equation or pseudocode block for easier replication.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on our manuscript. We will address each major comment below and revise the paper to incorporate the suggested improvements for clarity and robustness.

read point-by-point responses

  1. Referee: [Transferable item augmenter description] The central claim for the transferable item augmenter (described in the method section following the abstract) is that it addresses imbalance while introducing less irrelevant noise. This rests on the unverified assumption that LLM-generated cross-domain behaviors are domain-plausible and low-noise. No human evaluation, catalog-overlap metric, noise quantification, or ablation isolating the augmenter from dual-phase training and profiling is reported, making it impossible to confirm the noise-reduction benefit or rule out that gains derive from simply adding more data.

    Authors: We appreciate this observation. Although the overall performance improvements suggest the effectiveness of our approach, we agree that direct evidence for the low-noise property of the generated behaviors would strengthen the paper. In the revised manuscript, we will add a catalog-overlap metric to measure domain plausibility of the generated items, include noise quantification by comparing generated behaviors to real interactions, and provide a dedicated ablation study isolating the transferable item augmenter while keeping the other components fixed. This will help demonstrate that the benefits are not solely due to increased data volume. revision: yes

  2. Referee: [Experiments and results] The experimental validation (results section) reports effectiveness on three datasets but provides no error bars, statistical significance tests, or full ablation tables isolating each module. Without these, the cross-domain performance claims cannot be assessed for robustness against the reader's noted concerns about post-hoc choices or data leakage.

    Authors: We acknowledge the importance of these statistical validations. In the revised version, we will include error bars representing standard deviations from multiple random seeds, perform statistical significance tests (e.g., paired t-tests) between our method and baselines, and expand the ablation tables to isolate the contribution of each module individually. Additionally, we will provide more details on the experimental setup to address potential concerns about post-hoc choices and explicitly confirm the absence of data leakage in our cross-domain splits. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical method with external validation

full rationale

The paper introduces LLM-EDT as an architectural proposal consisting of a transferable item augmenter, dual-phase training strategy, and domain-aware profiling module. These components are motivated by identified CDSR challenges and evaluated via experiments on three public datasets. No mathematical derivation chain, fitted-parameter predictions, or load-bearing self-citations are present in the provided text that would reduce claimed outcomes to inputs by construction. The work relies on empirical results against external benchmarks rather than internal redefinitions or ansatzes.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The method rests on the assumption that LLMs possess sufficient world knowledge to generate useful cross-domain behaviors and that dual-phase training can separate shared and domain-specific signals without explicit supervision for the separation.

axioms (2)
  • domain assumption Large language models can act as reliable generators of plausible user-item interactions across domains without introducing dominant noise.
    Invoked when proposing the transferable item augmenter to address imbalance while claiming less irrelevant noise.
  • domain assumption Dual-phase training can empower domain-specific threads with domain-shared background knowledge.
    Stated as the mechanism to alleviate the transition issue in mixed interaction sequences.

pith-pipeline@v0.9.0 · 5851 in / 1472 out tokens · 27104 ms · 2026-05-21T18:51:24.646179+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 1 Pith paper

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

  1. Bridging Behavior and Semantics for Time-aware Cross-Domain Sequential Recommendation

    cs.IR 2026-05 unverdicted novelty 6.0

    BST-CDSR combines neural ODEs for continuous behavioral preference modeling with LLM-based temporal semantic generation and adaptive domain transfer to improve cross-domain sequential recommendations.

Reference graph

Works this paper leans on

37 extracted references · 37 canonical work pages · cited by 1 Pith paper · 1 internal anchor

  1. [1]

    Mohiuddin Ahmed, Raihan Seraj, and Syed Mohammed Shamsul Islam. 2020. The k-means algorithm: A comprehensive survey and performance evaluation. Electronics9, 8 (2020), 1295

    work page 2020

  2. [2]

    Jiangxia Cao, Xin Cong, Jiawei Sheng, Tingwen Liu, and Bin Wang. 2022. Con- trastive cross-domain sequential recommendation. InProceedings of the 31st ACM International Conference on Information & Knowledge Management. 138–147

    work page 2022

  3. [3]

    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)

    work page arXiv 2024

  4. [4]

    Yujia Ding, Huan Li, Ke Chen, and Lidan Shou. 2023. Tpuf: Enhancing cross- domain sequential recommendation via transferring pre-trained user features. In Proceedings of the 32nd ACM international conference on information and knowl- edge management. 410–419

    work page 2023

  5. [5]

    Jingtong Gao, Zhaocheng Du, Xiaopeng Li, Yichao Wang, Xiangyang Li, Huifeng Guo, Ruiming Tang, and Xiangyu Zhao. 2025. SampleLLM: Optimizing Tabular Data Synthesis in Recommendations. InCompanion Proceedings of the ACM on Web Conference 2025. 211–220

    work page 2025

  6. [6]

    Lei Guo, Li Tang, Tong Chen, Lei Zhu, Quoc Viet Hung Nguyen, and Hongzhi Yin. 2021. DA-GCN: A domain-aware attentive graph convolution network for shared-account cross-domain sequential recommendation.arXiv preprint arXiv:2105.03300(2021)

    work page arXiv 2021

  7. [7]

    Lei Guo, Hongzhi Yin, Qinyong Wang, Tong Chen, Alexander Zhou, and Nguyen Quoc Viet Hung. 2019. Streaming session-based recommendation. InProceedings of the 25th ACM SIGKDD international conference on knowledge discovery & data mining. 1569–1577

    work page 2019

  8. [8]

    Jun Hu, Wenwen Xia, Xiaolu Zhang, Chilin Fu, Weichang Wu, Zhaoxin Huan, Ang Li, Zuoli Tang, and Jun Zhou. 2024. Enhancing sequential recommendation via llm-based semantic embedding learning. InCompanion Proceedings of the ACM Web Conference 2024. 103–111

    work page 2024

  9. [9]

    Feiran Huang, Zhenghang Yang, Junyi Jiang, Yuanchen Bei, Yijie Zhang, and Hao Chen. 2024. Large language model interaction simulator for cold-start item recommendation.CoRR(2024)

    work page 2024

  10. [10]

    Dietmar Jannach and Malte Ludewig. 2017. When recurrent neural networks meet the neighborhood for session-based recommendation. InProceedings of the eleventh ACM conference on recommender systems. 306–310

    work page 2017

  11. [11]

    Yang Jiao, Fan Yang, Yetian Chen, Yan Gao, Jia Liu, and Yi Sun. 2024. Rethinking se- quential relationships: Improving sequential recommenders with inter-sequence data augmentation. InCompanion Proceedings of the ACM Web Conference 2024. 641–645

    work page 2024

  12. [12]

    Wang-Cheng Kang and Julian McAuley. 2018. Self-attentive sequential recom- mendation. In2018 IEEE international conference on data mining (ICDM). IEEE, 197–206

    work page 2018

  13. [13]

    Pan Li and Alexander Tuzhilin. 2020. Ddtcdr: Deep dual transfer cross domain recommendation. InProceedings of the 13th international conference on web search and data mining. 331–339

    work page 2020

  14. [14]

    Qidong Liu, Xian Wu, Wanyu Wang, Yejing Wang, Yuanshao Zhu, Xiangyu Zhao, Feng Tian, and Yefeng Zheng. 2025. Llmemb: Large language model can be a good embedding generator for sequential recommendation. InProceedings of the AAAI Conference on Artificial Intelligence, Vol. 39. 12183–12191

    work page 2025

  15. [15]

    Qidong Liu, Xian Wu, Yejing Wang, Zijian Zhang, Feng Tian, Yefeng Zheng, and Xiangyu Zhao. 2024. Llm-esr: Large language models enhancement for long- tailed sequential recommendation.Advances in Neural Information Processing Systems37 (2024), 26701–26727

    work page 2024

  16. [16]

    Qidong Liu, Xiangyu Zhao, Yuhao Wang, Yejing Wang, Zijian Zhang, Yuqi Sun, Xi- ang Li, Maolin Wang, Pengyue Jia, Chong Chen, et al. 2024. Large Language Model Enhanced Recommender Systems: A Survey.arXiv preprint arXiv:2412.13432 (2024)

    work page arXiv 2024

  17. [17]

    Qidong Liu, Xiangyu Zhao, Yejing Wang, Zijian Zhang, Howard Zhong, Chong Chen, Xiang Li, Wei Huang, and Feng Tian. 2025. Bridge the Domains: Large Language Models Enhanced Cross-domain Sequential Recommendation.arXiv preprint arXiv:2504.18383(2025)

    work page arXiv 2025

  18. [18]

    Haokai Ma, Ruobing Xie, Lei Meng, Xin Chen, Xu Zhang, Leyu Lin, and Jie Zhou. 2024. Triple sequence learning for cross-domain recommendation.ACM Transactions on Information Systems42, 4 (2024), 1–29

    work page 2024

  19. [19]

    Chung Park, Taesan Kim, Hyungjun Yoon, Junui Hong, Yelim Yu, Mincheol Cho, Minsung Choi, and Jaegul Choo. 2024. Pacer and runner: Cooperative learning framework between single-and cross-domain sequential recommendation. In Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval. 2071–2080

    work page 2024

  20. [20]

    Aske Plaat, Annie Wong, Suzan Verberne, Joost Broekens, Niki van Stein, and Thomas Back. 2024. Reasoning with large language models, a survey.arXiv preprint arXiv:2407.11511(2024)

    work page arXiv 2024

  21. [21]

    Xubin Ren, Wei Wei, Lianghao Xia, Lixin Su, Suqi Cheng, Junfeng Wang, Dawei Yin, and Chao Huang. 2024. Representation learning with large language models for recommendation. InProceedings of the ACM web conference 2024. 3464–3475

    work page 2024

  22. [22]

    Franco Scarselli, Marco Gori, Ah Chung Tsoi, Markus Hagenbuchner, and Gabriele Monfardini. 2008. The graph neural network model.IEEE transactions on neural networks20, 1 (2008), 61–80

    work page 2008

  23. [23]

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

    work page

  24. [24]

    InProceedings of the 28th ACM international conference on information and knowledge management

    BERT4Rec: Sequential recommendation with bidirectional encoder rep- resentations from transformer. InProceedings of the 28th ACM international conference on information and knowledge management. 1441–1450

    work page

  25. [25]

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need.Advances in neural information processing systems30 (2017)

    work page 2017

  26. [26]

    Junting Wang, Praneet Rathi, and Hari Sundaram. 2024. A pre-trained zero-shot sequential recommendation framework via popularity dynamics. InProceedings of the 18th ACM Conference on Recommender Systems. 433–443

    work page 2024

  27. [27]

    Shoujin Wang, Liang Hu, Yan Wang, Longbing Cao, Quan Z Sheng, and Mehmet Orgun. 2019. Sequential recommender systems: challenges, progress and prospects.arXiv preprint arXiv:2001.04830(2019)

    work page arXiv 2019

  28. [28]

    Xu Xie, Fei Sun, Zhaoyang Liu, Shiwen Wu, Jinyang Gao, Jiandong Zhang, Bolin Ding, and Bin Cui. 2022. Contrastive learning for sequential recommendation. In 2022 IEEE 38th international conference on data engineering (ICDE). IEEE, 1259– 1273

    work page 2022

  29. [29]

    Haoran Xin, Ying Sun, Chao Wang, and Hui Xiong. 2025. Llmcdsr: Enhancing cross-domain sequential recommendation with large language models.ACM Transactions on Information Systems(2025)

    work page 2025

  30. [30]

    Wujiang Xu, Qitian Wu, Runzhong Wang, Mingming Ha, Qiongxu Ma, Linxun Chen, Bing Han, and Junchi Yan. 2024. Rethinking cross-domain sequential recommendation under open-world assumptions. InProceedings of the ACM Web Conference 2024. 3173–3184

    work page 2024

  31. [31]

    Jia-Yu Yao, Kun-Peng Ning, Zhen-Hui Liu, Mu-Nan Ning, Yu-Yang Liu, and Li Yuan. 2023. Llm lies: Hallucinations are not bugs, but features as adversarial examples.arXiv preprint arXiv:2310.01469(2023)

    work page arXiv 2023

  32. [32]

    Tianzi Zang, Yanmin Zhu, Haobing Liu, Ruohan Zhang, and Jiadi Yu. 2022. A survey on cross-domain recommendation: taxonomies, methods, and future directions.ACM Transactions on Information Systems41, 2 (2022), 1–39

    work page 2022

  33. [33]

    Peitian Zhang, Zheng Liu, Shitao Xiao, Zhicheng Dou, and Jian-Yun Nie. 2024. A multi-task embedder for retrieval augmented LLMs. InProceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 3537–3553

    work page 2024

  34. [34]

    Wayne Xin Zhao, Kun Zhou, Junyi Li, Tianyi Tang, Xiaolei Wang, Yupeng Hou, Yingqian Min, Beichen Zhang, Junjie Zhang, Zican Dong, et al. 2023. A survey of large language models.arXiv preprint arXiv:2303.182231, 2 (2023)

    work page internal anchor Pith review Pith/arXiv arXiv 2023

  35. [35]

    Guorui Zhou, Na Mou, Ying Fan, Qi Pi, Weijie Bian, Chang Zhou, Xiaoqiang Zhu, and Kun Gai. 2019. Deep interest evolution network for click-through rate prediction. InProceedings of the AAAI conference on artificial intelligence, Vol. 33. 5941–5948

    work page 2019

  36. [36]

    Guorui Zhou, Xiaoqiang Zhu, Chenru Song, Ying Fan, Han Zhu, Xiao Ma, Yanghui Yan, Junqi Jin, Han Li, and Kun Gai. 2018. Deep interest network for click-through rate prediction. InProceedings of the 24th ACM SIGKDD international conference on knowledge discovery & data mining. 1059–1068

    work page 2018

  37. [37]

    pseudo items

    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. InProceedings of the fifteenth ACM international conference on web search and data mining. 1507–1515. A Setting and Descriptions A.1 Prompt Templates In this section, we will p...

    work page 2022