pith. sign in

arxiv: 2605.29141 · v1 · pith:7AHUXNQ3new · submitted 2026-05-27 · 💻 cs.IR · cs.AI

Toward User Preference Alignment in LLM Recommendation via Explicit Context Feedback

Weizhi Zhang , Wooseong Yang , Yuxin Cui , Zhaohui Guo , Hins Hu , Liangwei Yang , Henry Peng Zou , Qifei Wang
show 4 more authors
Hanqing Zeng Jiayi Liu Yinglong Xia Philip S. Yu
This is my paper

Pith reviewed 2026-06-29 09:28 UTC · model grok-4.3

classification 💻 cs.IR cs.AI
keywords recommender systemslarge language modelsexplicit feedbackuser preferencescontext feedbackexplainable AIpreference alignment
0
0 comments X

The pith

LLM recommender systems should prioritize explicit context feedback from user comments to achieve better preference alignment.

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

The paper claims that current LLM-based recommender systems underutilize explicit context feedback such as user comments and reviews, which provide the semantic reasons for preferences. Instead, they rely on implicit signals and item metadata, leading to misaligned recommendations and filter bubbles. By integrating this feedback, systems can produce more accurate, diverse, and explainable results. The authors review recommendation paradigms, value the context-rich signals, and outline frameworks and benchmarks for future LLM RecSys.

Core claim

Explicit context feedback captures nuanced reasons behind user decisions and offers heterogeneous information for preference alignment and explainable recommendations. Overlooking it misaligns preferences. Recent LLM advances allow harnessing user-generated content, but current systems still focus on metadata. The paper advocates prioritizing this feedback in next-generation LLM-based RecSys through new frameworks for integration.

What carries the argument

Frameworks for integrating explicit user signals into scalable LLM-driven RecSys, which use context-rich feedback to model user preferences.

If this is right

  • More accurate recommendations by understanding semantic context behind choices.
  • Reduced filter bubbles through better preference alignment.
  • Enhanced explainability of why certain items are recommended.
  • Development of new benchmarks and metrics focused on context feedback.
  • Scalable integration of user text signals into LLM systems.

Where Pith is reading between the lines

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

  • Platforms might need to redesign interfaces to collect more explicit feedback.
  • This approach could generalize to other AI systems that model user intent from text.
  • Potential to address biases in recommendations by incorporating diverse user expressions.

Load-bearing premise

Explicit context feedback is underutilized and LLMs can harness it effectively for preference alignment without introducing scalability or bias issues.

What would settle it

Comparing an LLM recommender using only implicit data against one that also processes explicit review text, measuring differences in accuracy, diversity, and explanation quality.

Figures

Figures reproduced from arXiv: 2605.29141 by Hanqing Zeng, Henry Peng Zou, Hins Hu, Jiayi Liu, Liangwei Yang, Philip S. Yu, Qifei Wang, Weizhi Zhang, Wooseong Yang, Yinglong Xia, Yuxin Cui, Zhaohui Guo.

Figure 1
Figure 1. Figure 1: Limitations of existing recommender systems (primarily rely on implicit feedback while overlooking explicit contextual feedback) and framework [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
read the original abstract

Traditional recommender systems (RecSys) primarily infer user preferences from implicit signals (such as clicks, watches, and purchases), often neglecting the rich explicit contextual feedback users provide through verbal text, like comments and reviews. This explicit context feedback captures the nuanced reasons behind user decisions regarding their preferences. In addition, it offers critical heterogeneous information for user preference alignment and more explainable recommendations. Overlooking such signals can lead to misaligned user preferences and further reinforce filter bubbles, as algorithms fail to understand the "semantic context" behind user choices. Recent advances in Large Language Models (LLMs) present new opportunities to harness user-generated content for more accurate and diverse recommendations, yet current LLM-based recommendations still focus on using item meta-data and underutilize this resource. In this paper, we advocate for prioritizing explicit context feedback in the next generation of LLM-based RecSys. We review the evolution of recommendation paradigms, highlight the value of context-rich feedback, call for new benchmarks and metrics, and introduce frameworks for integrating explicit user signals into scalable LLM-driven RecSys. Centering on user-preference modeling, we aim to foster more personalized, transparent, and explainable RecSys online platforms.

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 / 1 minor

Summary. The manuscript is a position paper advocating prioritization of explicit context feedback (user comments, reviews, and verbal text) in LLM-based recommender systems. It contrasts this with traditional implicit-signal RecSys and current LLM approaches that rely primarily on item metadata, arguing that explicit signals capture semantic reasons for preferences, enable better alignment, diversity, and explainability, and help avoid filter bubbles. The paper reviews the evolution of recommendation paradigms, highlights the value of context-rich feedback, calls for new benchmarks and metrics, and introduces high-level frameworks for integrating these signals into scalable LLM-driven systems.

Significance. If the advocated shift proves effective, the work could stimulate research on semantically richer user modeling in RecSys, addressing limitations in personalization and transparency. The call for new benchmarks targets an evaluation gap in context-aware LLM recommendations. As a position paper with frameworks, it provides a conceptual roadmap rather than empirical results.

major comments (2)
  1. [Abstract] Abstract: The central motivation—that 'current LLM-based recommendations still focus on using item meta-data and underutilize this resource'—is stated without citing specific prior LLM-RecSys works or providing even a brief survey of current practices. This undercuts the load-bearing claim that explicit context feedback is a neglected opportunity.
  2. The introduced frameworks for integrating explicit user signals are described only at a conceptual level with no discussion of implementation challenges (e.g., handling noisy or sparse textual feedback at scale, or potential introduction of new biases from LLM processing of reviews). This weakens the claim of 'scalable LLM-driven RecSys' without concrete grounding.
minor comments (1)
  1. The review of recommendation paradigms would benefit from explicit section headings or a timeline figure to improve readability for readers unfamiliar with the historical arc.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the thoughtful review and constructive suggestions. As a position paper, our goal is to outline a conceptual direction rather than provide exhaustive empirical validation. We address each major comment below and will incorporate revisions to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The central motivation—that 'current LLM-based recommendations still focus on using item meta-data and underutilize this resource'—is stated without citing specific prior LLM-RecSys works or providing even a brief survey of current practices. This undercuts the load-bearing claim that explicit context feedback is a neglected opportunity.

    Authors: We agree that the abstract would benefit from explicit citations to support the claim. In the revised manuscript, we will add representative references to recent LLM-RecSys works that primarily rely on item metadata (such as those focusing on sequential recommendation or metadata-augmented prompting) and include a concise survey of prevailing practices in the introduction section to better ground the motivation. revision: yes

  2. Referee: The introduced frameworks for integrating explicit user signals are described only at a conceptual level with no discussion of implementation challenges (e.g., handling noisy or sparse textual feedback at scale, or potential introduction of new biases from LLM processing of reviews). This weakens the claim of 'scalable LLM-driven RecSys' without concrete grounding.

    Authors: The frameworks are presented at a high level consistent with the position-paper format, which prioritizes vision over detailed engineering. However, we acknowledge that briefly addressing implementation realities would improve the discussion of scalability. We will add a short paragraph in the frameworks section noting key challenges, including noise and sparsity in textual feedback, computational scaling considerations, and potential biases introduced by LLM interpretation of reviews. revision: yes

Circularity Check

0 steps flagged

Position paper with no derivation chain

full rationale

This is a position paper that reviews prior paradigms, argues for prioritizing explicit context feedback in LLM-based RecSys, and outlines high-level frameworks plus calls for new benchmarks. It contains no equations, no fitted parameters, no predictions of quantities, and no derivation steps that could reduce to inputs by construction or via self-citation chains. All central claims are aspirational advocacy rather than technical assertions whose correctness depends on an internal chain that loops back to the paper's own definitions or fits.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

No technical derivations or new constructs are detailed in the abstract; the contribution is conceptual advocacy.

pith-pipeline@v0.9.1-grok · 5778 in / 866 out tokens · 22079 ms · 2026-06-29T09:28:46.939702+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

80 extracted references · 19 canonical work pages · 6 internal anchors

  1. [1]

    Recommender systems based on user reviews: the state of the art,

    L. Chen, G. Chen, and F. Wang, “Recommender systems based on user reviews: the state of the art,”User Modeling and User-Adapted Interaction, vol. 25, no. 2, pp. 99–154, 2015

    2015

  2. [2]

    Based recommender systems: a survey of approaches, challenges and future perspectives,

    E. Hasan, M. Rahman, C. Ding, J. Huang, and S. Raza, “Based recommender systems: a survey of approaches, challenges and future perspectives,”ACM Computing Surveys, 2024

    2024

  3. [3]

    A review-aware graph contrastive learning framework for recommenda- tion,

    J. Shuai, K. Zhang, L. Wu, P. Sun, R. Hong, M. Wang, and Y . Li, “A review-aware graph contrastive learning framework for recommenda- tion,” inProceedings of the 45th international ACM SIGIR conference on research and development in information retrieval, 2022, pp. 1283– 1293

    2022

  4. [4]

    Counterfactual review-based recommendation,

    K. Xiong, W. Ye, X. Chen, Y . Zhang, W. X. Zhao, B. Hu, Z. Zhang, and J. Zhou, “Counterfactual review-based recommendation,” inProceedings of the 30th ACM International Conference on Information & Knowledge Management, 2021, pp. 2231–2240

    2021

  5. [5]

    Beyond explicit and implicit: How users provide feedback to shape personalized recommen- dation content,

    W. Li, J.-C. Kuo, M. Sheng, P. Zhang, and Q. Wu, “Beyond explicit and implicit: How users provide feedback to shape personalized recommen- dation content,” inProceedings of the 2025 CHI Conference on Human Factors in Computing Systems, 2025, pp. 1–17

    2025

  6. [6]

    OneRec: Unifying Retrieve and Rank with Generative Recommender and Iterative Preference Alignment

    J. Deng, S. Wang, K. Cai, L. Ren, Q. Hu, W. Ding, Q. Luo, and G. Zhou, “Onerec: Unifying retrieve and rank with generative recommender and iterative preference alignment,”arXiv preprint arXiv:2502.18965, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  7. [7]

    Tutorial on large language models for recommendation,

    W. Hua, L. Li, S. Xu, L. Chen, and Y . Zhang, “Tutorial on large language models for recommendation,” inProceedings of the 17th ACM Conference on Recommender Systems, 2023, pp. 1281–1283

    2023

  8. [8]

    LLM-based User Profile Management for Recommender System

    S. Bang and H. Song, “Llm-based user profile management for recom- mender system,”arXiv preprint arXiv:2502.14541, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  9. [9]

    User feedback alignment for llm-powered exploration in large-scale recommendation systems,

    J. Wang, Y . Liu, Y . Sun, X. Ma, Y . Wang, H. Ma, Z. Su, M. Chen, M. Gao, O. Dalalet al., “User feedback alignment for llm-powered exploration in large-scale recommendation systems,”arXiv preprint arXiv:2504.05522, 2025

    work page arXiv 2025

  10. [10]

    Content-based recommendation systems,

    M. J. Pazzani and D. Billsus, “Content-based recommendation systems,” inThe adaptive web: methods and strategies of web personalization. Springer, 2007, pp. 325–341

    2007

  11. [11]

    A review of content-based and context-based recommendation systems,

    U. Javed, K. Shaukat, I. A. Hameed, F. Iqbal, T. M. Alam, and S. Luo, “A review of content-based and context-based recommendation systems,” International Journal of Emerging Technologies in Learning (iJET), vol. 16, no. 3, pp. 274–306, 2021

    2021

  12. [12]

    Wide & deep learning for recommender systems,

    H.-T. Cheng, L. Koc, J. Harmsen, T. Shaked, T. Chandra, H. Aradhye, G. Anderson, G. Corrado, W. Chai, M. Ispiret al., “Wide & deep learning for recommender systems,” inProceedings of the 1st workshop on deep learning for recommender systems, 2016, pp. 7–10

    2016

  13. [13]

    Deepfm: a factorization- machine based neural network for ctr prediction,

    H. Guo, R. Tang, Y . Ye, Z. Li, and X. He, “Deepfm: a factorization- machine based neural network for ctr prediction,” inProceedings of the 26th International Joint Conference on Artificial Intelligence, 2017, pp. 1725–1731

    2017

  14. [14]

    Deep & cross network for ad click predictions,

    R. Wang, B. Fu, G. Fu, and M. Wang, “Deep & cross network for ad click predictions,” inProceedings of the ADKDD’17, 2017, pp. 1–7

    2017

  15. [15]

    Learning deep structured semantic models for web search using clickthrough data,

    P.-S. Huang, X. He, J. Gao, L. Deng, A. Acero, and L. Heck, “Learning deep structured semantic models for web search using clickthrough data,” inProceedings of the 22nd ACM international conference on Information & Knowledge Management, 2013, pp. 2333–2338

    2013

  16. [16]

    Cold-start recommendation towards the era of large language models (llms): A comprehensive survey and roadmap,

    W. Zhang, Y . Bei, L. Yang, H. P. Zou, P. Zhou, A. Liu, Y . Li, H. Chen, J. Wang, Y . Wanget al., “Cold-start recommendation towards the era of large language models (llms): A comprehensive survey and roadmap,” arXiv preprint arXiv:2501.01945, 2025

    work page arXiv 2025

  17. [17]

    Diversity, serendipity, novelty, and coverage: a survey and empirical analysis of beyond-accuracy objectives in recommender systems,

    M. Kaminskas and D. Bridge, “Diversity, serendipity, novelty, and coverage: a survey and empirical analysis of beyond-accuracy objectives in recommender systems,”ACM Transactions on Interactive Intelligent Systems (TiiS), vol. 7, no. 1, pp. 1–42, 2016

    2016

  18. [18]

    Echo chambers, filter bubbles, and polarisation: A literature review,

    A. Ross Arguedas, C. Robertson, R. Fletcher, and R. Nielsen, “Echo chambers, filter bubbles, and polarisation: A literature review,” 2022

    2022

  19. [19]

    Advances in collaborative filtering,

    Y . Koren, S. Rendle, and R. Bell, “Advances in collaborative filtering,” Recommender systems handbook, pp. 91–142, 2021

    2021

  20. [20]

    Mixed supervised graph contrastive learning for recommendation,

    W. Zhang, L. Yang, Z. Song, H. P. Zou, K. Xu, Y . Zhu, and P. S. Yu, “Mixed supervised graph contrastive learning for recommendation,” arXiv preprint arXiv:2404.15954, 2024

    work page arXiv 2024

  21. [21]

    A collaborative filtering recommender systems: Survey,

    M. F. Aljunid, D. Manjaiah, M. K. Hooshmand, W. A. Ali, A. M. Shetty, and S. Q. Alzoubah, “A collaborative filtering recommender systems: Survey,”Neurocomputing, vol. 617, p. 128718, 2025

    2025

  22. [22]

    Prob- abilistic memory-based collaborative filtering,

    K. Yu, A. Schwaighofer, V . Tresp, X. Xu, and H.-P. Kriegel, “Prob- abilistic memory-based collaborative filtering,”IEEE Transactions on Knowledge and Data Engineering, vol. 16, no. 1, pp. 56–69, 2004

    2004

  23. [23]

    Model-based collaborative filtering,

    C. C. Aggarwal, “Model-based collaborative filtering,” inRecommender systems: the textbook. Springer, 2016, pp. 71–138

    2016

  24. [24]

    Matrix factorization model in collaborative filtering algorithms: A survey,

    D. Bokde, S. Girase, and D. Mukhopadhyay, “Matrix factorization model in collaborative filtering algorithms: A survey,”Procedia Computer Science, vol. 49, pp. 136–146, 2015

    2015

  25. [25]

    Neural col- laborative filtering,

    X. He, L. Liao, H. Zhang, L. Nie, X. Hu, and T.-S. Chua, “Neural col- laborative filtering,” inProceedings of the 26th international conference on world wide web, 2017, pp. 173–182

    2017

  26. [26]

    A comprehensive survey on graph neural networks,

    Z. Wu, S. Pan, F. Chen, G. Long, C. Zhang, and P. S. Yu, “A comprehensive survey on graph neural networks,”IEEE transactions on neural networks and learning systems, vol. 32, no. 1, pp. 4–24, 2020

    2020

  27. [27]

    Graph-based village level poverty identification,

    J. Ma, L. Yang, Q. Feng, W. Zhang, and P. S. Yu, “Graph-based village level poverty identification,” inProceedings of the ACM web conference 2023, 2023, pp. 4115–4119

    2023

  28. [28]

    Lightgcn: Simplifying and powering graph convolution network for recommenda- tion,

    X. He, K. Deng, X. Wang, Y . Li, Y . Zhang, and M. Wang, “Lightgcn: Simplifying and powering graph convolution network for recommenda- tion,” inProceedings of the 43rd International ACM SIGIR conference on research and development in Information Retrieval, 2020, pp. 639– 648

    2020

  29. [29]

    Self- supervised graph learning for recommendation,

    J. Wu, X. Wang, F. Feng, X. He, L. Chen, J. Lian, and X. Xie, “Self- supervised graph learning for recommendation,” inProceedings of the 44th international ACM SIGIR conference on research and development in information retrieval, 2021, pp. 726–735

    2021

  30. [30]

    Are graph augmentations necessary? simple graph contrastive learning for recommendation,

    J. Yu, H. Yin, X. Xia, T. Chen, L. Cui, and Q. V . H. Nguyen, “Are graph augmentations necessary? simple graph contrastive learning for recommendation,” inProceedings of the 45th international ACM SIGIR conference on research and development in information retrieval, 2022, pp. 1294–1303

    2022

  31. [31]

    Sgcl: Unifying self-supervised and supervised learning for graph recommendation,

    W. Zhang, L. Yang, Z. Song, H. P. Zou, K. Xu, Y . Zhu, and P. S. Yu, “Sgcl: Unifying self-supervised and supervised learning for graph recommendation,” inProceedings of the Nineteenth ACM Conference on Recommender Systems, 2025, pp. 671–676

    2025

  32. [32]

    Ultragcn: ultra simplification of graph convolutional networks for recommendation,

    K. Mao, J. Zhu, X. Xiao, B. Lu, Z. Wang, and X. He, “Ultragcn: ultra simplification of graph convolutional networks for recommendation,” in Proceedings of the 30th ACM international conference on information & knowledge management, 2021, pp. 1253–1262

    2021

  33. [33]

    Do we really need graph convolution during training? light post- training graph-ode for efficient recommendation,

    W. Zhang, L. Yang, Z. Song, H. P. Zou, K. Xu, L. Fang, and P. S. Yu, “Do we really need graph convolution during training? light post- training graph-ode for efficient recommendation,” inProceedings of the 33rd ACM International Conference on Information and Knowledge Management, 2024, pp. 3248–3258

    2024

  34. [34]

    Graph neural ordinary differential equations-based method for collaborative filtering,

    K. Xu, Y . Zhu, W. Zhang, and P. S. Yu, “Graph neural ordinary differential equations-based method for collaborative filtering,” in2023 IEEE International Conference on Data Mining (ICDM). IEEE, 2023, pp. 1445–1450

    2023

  35. [35]

    Graph neural con- trolled differential equations for collaborative filtering,

    K. Xu, W. Zhang, Z. Song, Y . Zhu, and P. S. Yu, “Graph neural con- trolled differential equations for collaborative filtering,” inCompanion Proceedings of the ACM on Web Conference 2025, 2025, pp. 1446–1449

    2025

  36. [36]

    Bias and debias in recommender system: A survey and future directions,

    J. Chen, H. Dong, X. Wang, F. Feng, M. Wang, and X. He, “Bias and debias in recommender system: A survey and future directions,”ACM Transactions on Information Systems, vol. 41, no. 3, pp. 1–39, 2023

    2023

  37. [37]

    A survey on popularity bias in recommender systems,

    A. Klimashevskaia, D. Jannach, M. Elahi, and C. Trattner, “A survey on popularity bias in recommender systems,”User Modeling and User- Adapted Interaction, vol. 34, no. 5, pp. 1777–1834, 2024

    2024

  38. [38]

    Sequential recommender systems: challenges, progress and prospects,

    S. Wang, L. Hu, Y . Wang, L. Cao, Q. Z. Sheng, and M. Orgun, “Sequential recommender systems: challenges, progress and prospects,” arXiv preprint arXiv:2001.04830, 2019

    work page arXiv 2001

  39. [39]

    Factorizing personalized markov chains for next-basket recommendation,

    S. Rendle, C. Freudenthaler, and L. Schmidt-Thieme, “Factorizing personalized markov chains for next-basket recommendation,” inPro- ceedings of the 19th international conference on World wide web, 2010, pp. 811–820

    2010

  40. [40]

    Session-based Recommendations with Recurrent Neural Networks

    B. Hidasi, A. Karatzoglou, L. Baltrunas, and D. Tikk, “Session- based recommendations with recurrent neural networks,”arXiv preprint arXiv:1511.06939, 2015

    work page internal anchor Pith review Pith/arXiv arXiv 2015

  41. [41]

    Self-attentive sequential recommenda- tion,

    W.-C. Kang and J. McAuley, “Self-attentive sequential recommenda- tion,” in2018 IEEE international conference on data mining (ICDM). IEEE, 2018, pp. 197–206

    2018

  42. [42]

    Bert4rec: Sequential recommendation with bidirectional encoder representations from transformer,

    F. Sun, J. Liu, J. Wu, C. Pei, X. Lin, W. Ou, and P. Jiang, “Bert4rec: Sequential recommendation with bidirectional encoder representations from transformer,” inProceedings of the 28th ACM international confer- ence on information and knowledge management, 2019, pp. 1441–1450

    2019

  43. [43]

    Recommender systems in the era of large language models (llms),

    Z. Zhao, W. Fan, J. Li, Y . Liu, X. Mei, Y . Wang, Z. Wen, F. Wang, X. Zhao, J. Tanget al., “Recommender systems in the era of large language models (llms),”IEEE Transactions on Knowledge and Data Engineering, vol. 36, no. 11, pp. 6889–6907, 2024

    2024

  44. [44]

    Text is all you need: Learning language representations for sequential recommendation,

    J. Li, M. Wang, J. Li, J. Fu, X. Shen, J. Shang, and J. McAuley, “Text is all you need: Learning language representations for sequential recommendation,” inProceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2023, pp. 1258–1267

    2023

  45. [45]

    Recommendation as language processing (rlp): A unified pretrain, personalized prompt & predict paradigm (p5),

    S. Geng, S. Liu, Z. Fu, Y . Ge, and Y . Zhang, “Recommendation as language processing (rlp): A unified pretrain, personalized prompt & predict paradigm (p5),” inProceedings of the 16th ACM conference on recommender systems, 2022, pp. 299–315

    2022

  46. [46]

    Large language models are zero-shot rankers for recommender systems,

    Y . Hou, J. Zhang, Z. Lin, H. Lu, R. Xie, J. McAuley, and W. X. Zhao, “Large language models are zero-shot rankers for recommender systems,” inEuropean Conference on Information Retrieval. Springer, 2024, pp. 364–381

    2024

  47. [47]

    The movielens datasets: History and context,

    F. M. Harper and J. A. Konstan, “The movielens datasets: History and context,”Acm transactions on interactive intelligent systems (tiis), vol. 5, no. 4, pp. 1–19, 2015

    2015

  48. [48]

    Bridging Language and Items for Retrieval and Recommendation: Benchmarking LLMs as Semantic Encoders

    Y . Hou, J. Li, Z. He, A. Yan, X. Chen, and J. McAuley, “Bridging language and items for retrieval and recommendation,”arXiv preprint arXiv:2403.03952, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  49. [49]

    Improving aggregate recommendation diversity using ranking-based techniques,

    G. Adomavicius and Y . Kwon, “Improving aggregate recommendation diversity using ranking-based techniques,”IEEE Transactions on Knowl- edge and Data Engineering, vol. 24, no. 5, pp. 896–911, 2011

    2011

  50. [50]

    Novelty and diversity in top-n recommendation–analysis and evaluation,

    N. Hurley and M. Zhang, “Novelty and diversity in top-n recommendation–analysis and evaluation,”ACM Transactions on Inter- net Technology (TOIT), vol. 10, no. 4, pp. 1–30, 2011

    2011

  51. [51]

    Joint deep modeling of users and items using reviews for recommendation,

    L. Zheng, V . Noroozi, and P. S. Yu, “Joint deep modeling of users and items using reviews for recommendation,” inProceedings of the tenth ACM international conference on web search and data mining, 2017, pp. 425–434

    2017

  52. [52]

    Neural attentional rating regression with review-level explanations,

    C. Chen, M. Zhang, Y . Liu, and S. Ma, “Neural attentional rating regression with review-level explanations,” inProceedings of the 2018 world wide web conference, 2018, pp. 1583–1592

    2018

  53. [53]

    A contrastive framework with user, item and review alignment for recommendation,

    H. V . Dong, Y . Fang, and H. W. Lauw, “A contrastive framework with user, item and review alignment for recommendation,” inProceedings of the Eighteenth ACM International Conference on Web Search and Data Mining, 2025, pp. 117–126

    2025

  54. [54]

    Item cluster-aware prompt learning for session-based recommendation,

    W. Yang, C. Wang, Z. Song, W. Zhang, and P. S. Yu, “Item cluster-aware prompt learning for session-based recommendation,”arXiv preprint arXiv:2410.04756, 2024

    work page arXiv 2024

  55. [55]

    Rdrec: Rationale distilla- tion for llm-based recommendation,

    X. Wang, J. Cui, Y . Suzuki, and F. Fukumoto, “Rdrec: Rationale distilla- tion for llm-based recommendation,”arXiv preprint arXiv:2405.10587, 2024

    work page arXiv 2024

  56. [56]

    Liu,Sentiment analysis and opinion mining

    B. Liu,Sentiment analysis and opinion mining. Springer Nature, 2022

    2022

  57. [57]

    Contrastive learning for cold-start recommendation,

    Y . Wei, X. Wang, Q. Li, L. Nie, Y . Li, X. Li, and T.-S. Chua, “Contrastive learning for cold-start recommendation,” inProceedings of the 29th ACM international conference on multimedia, 2021, pp. 5382–5390

    2021

  58. [58]

    Learning from negative user feedback and measuring responsiveness for sequential recommenders,

    Y . Wang, Y . Halpern, S. Chang, J. Feng, E. Y . Le, L. Li, X. Liang, M.-C. Huang, S. Li, A. Beutelet al., “Learning from negative user feedback and measuring responsiveness for sequential recommenders,” inProceedings of the 17th ACM Conference on Recommender Systems, 2023, pp. 1049–1053

    2023

  59. [59]

    Context-aware recommender sys- tems,

    G. Adomavicius and A. Tuzhilin, “Context-aware recommender sys- tems,” inRecommender systems handbook. Springer, 2010, pp. 217– 253

    2010

  60. [60]

    Memory and new controls for chatgpt,

    OpenAI, “Memory and new controls for chatgpt,” https://openai.com/index/memory-and-new-controls-for-chatgpt/, 2024, accessed: 2025-09-26

    2024

  61. [61]

    Claude introduces memory for teams at work,

    Anthropic, “Claude introduces memory for teams at work,” https://www.anthropic.com/news/memory, 2024, accessed: 2025-09-26

    2024

  62. [62]

    Human-inspired perspectives: A survey on ai long-term memory,

    L. Heet al., “Human-inspired perspectives: A survey on ai long-term memory,”arXiv preprint arXiv:2411.00489, 2024

    work page arXiv 2024

  63. [63]

    Ader: Adaptively distilled exemplar replay towards continual learning for session-based recommendation,

    F. Mi, X. Lin, and B. Faltings, “Ader: Adaptively distilled exemplar replay towards continual learning for session-based recommendation,” inProceedings of the 14th ACM Conference on Recommender Systems, 2020, pp. 408–413

    2020

  64. [64]

    Rating and aspect-based opinion graph embeddings for explainable recommendations,

    I. Cantador, A. Carvallo, and F. Diez, “Rating and aspect-based opinion graph embeddings for explainable recommendations,”arXiv preprint arXiv:2107.03385, 2021

    work page arXiv 2021

  65. [65]

    Askat: Aspect sentiment knowledge graph attention network for recommendation,

    Y . Cui, P. Zhou, H. Yu, P. Sun, H. Cao, and P. Yang, “Askat: Aspect sentiment knowledge graph attention network for recommendation,” Electronics, vol. 13, no. 1, p. 216, 2024

    2024

  66. [66]

    Disentangled graph contrastive learning for review-based recommendation,

    Y . Ren, H. Zhang, Q. Li, L. Fu, J. Ding, X. Cao, X. Wang, and C. Zhou, “Disentangled graph contrastive learning for review-based recommendation,”arXiv preprint arXiv:2209.01524, 2022

    work page arXiv 2022

  67. [67]

    Learning hierarchical review graph representations for recommendation,

    Y . Liu, S. Yang, Y . Zhang, C. Miao, Z. Nie, and J. Zhang, “Learning hierarchical review graph representations for recommendation,”IEEE Transactions on Knowledge and Data Engineering, vol. 35, no. 1, pp. 658–671, 2021

    2021

  68. [68]

    Personax: A recommendation agent oriented user modeling framework for long behavior sequence,

    Y . Shi, W. Xu, Z. Zhang, X. Zi, Q. Wu, and M. Xu, “Personax: A recommendation agent oriented user modeling framework for long behavior sequence,”arXiv preprint arXiv:2503.02398, 2025

    work page arXiv 2025

  69. [69]

    PersonaAgent: Bridging Memory and Action for Personalized LLM Agents

    W. Zhang, X. Zhang, C. Zhang, L. Yang, J. Shang, Z. Wei, H. P. Zou, Z. Huang, Z. Wang, Y . Gaoet al., “Personaagent: When large language model agents meet personalization at test time,”arXiv preprint arXiv:2506.06254, 2025

    work page internal anchor Pith review Pith/arXiv arXiv 2025

  70. [70]

    Knowledge graph self- supervised rationalization for recommendation,

    Y . Yang, C. Huang, L. Xia, and C. Huang, “Knowledge graph self- supervised rationalization for recommendation,” inProceedings of the 29th ACM SIGKDD conference on knowledge discovery and data mining, 2023, pp. 3046–3056

    2023

  71. [71]

    Llminit: A free lunch from large language models for selective initialization of recommendation,

    W. Zhang, L. Yang, W. Yang, H. P. Zou, Y . Liu, K. Xu, S. Medya, and P. S. Yu, “Llminit: A free lunch from large language models for selective initialization of recommendation,”arXiv preprint arXiv:2503.01814, 2025

    work page arXiv 2025

  72. [72]

    Graphs meet ai agents: Taxonomy, progress, and future opportunities,

    Y . Bei, W. Zhang, S. Wang, W. Chen, S. Zhou, H. Chen, Y . Li, J. Bu, S. Pan, Y . Yuet al., “Graphs meet ai agents: Taxonomy, progress, and future opportunities,”arXiv preprint arXiv:2506.18019, 2025

    work page arXiv 2025

  73. [73]

    Explainable recommendation: A survey and new perspectives,

    Y . Zhang, X. Chenet al., “Explainable recommendation: A survey and new perspectives,”Foundations and Trends® in Information Retrieval, vol. 14, no. 1, pp. 1–101, 2020

    2020

  74. [74]

    Retrieval- augmented generation for knowledge-intensive nlp tasks,

    P. Lewis, E. Perez, A. Piktus, F. Petroni, V . Karpukhin, N. Goyal, H. K ¨uttler, M. Lewis, W.-t. Yih, T. Rockt ¨aschelet al., “Retrieval- augmented generation for knowledge-intensive nlp tasks,”Advances in neural information processing systems, vol. 33, pp. 9459–9474, 2020

    2020

  75. [75]

    Cold- start recommendation with knowledge-guided retrieval-augmented gen- eration,

    W. Yang, W. Zhang, Y . Liu, Y . Han, Y . Wang, J. Lee, and P. S. Yu, “Cold- start recommendation with knowledge-guided retrieval-augmented gen- eration,”arXiv preprint arXiv:2505.20773, 2025

    work page arXiv 2025

  76. [76]

    Improving recommendation lists through topic diversification,

    C.-N. Ziegler, S. M. McNee, J. A. Konstan, and G. Lausen, “Improving recommendation lists through topic diversification,” inProceedings of the 14th international conference on World Wide Web, 2005, pp. 22–32

    2005

  77. [77]

    Hybrid recommender systems: Survey and experiments,

    R. Burke, “Hybrid recommender systems: Survey and experiments,” User modeling and user-adapted interaction, vol. 12, no. 4, pp. 331–370, 2002

    2002

  78. [78]

    A Survey on LLM-as-a-Judge

    J. Gu, X. Jiang, Z. Shi, H. Tan, X. Zhai, C. Xu, W. Li, Y . Shen, S. Ma, H. Liuet al., “A survey on llm-as-a-judge,”arXiv preprint arXiv:2411.15594, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  79. [79]

    Image-based recommendations on styles and substitutes,

    J. McAuley, C. Targett, Q. Shi, and A. Van Den Hengel, “Image-based recommendations on styles and substitutes,” inProceedings of the 38th international ACM SIGIR conference on research and development in information retrieval, 2015, pp. 43–52

    2015

  80. [80]

    Atten- tive collaborative filtering: Multimedia recommendation with item-and component-level attention,

    J. Chen, H. Zhang, X. He, L. Nie, W. Liu, and T.-S. Chua, “Atten- tive collaborative filtering: Multimedia recommendation with item-and component-level attention,” inProceedings of the 40th International ACM SIGIR conference on Research and Development in Information Retrieval, 2017, pp. 335–344

    2017