LLM-based User Profile Management for Recommender System

Hwanjun Song; Seunghwan Bang

arxiv: 2502.14541 · v3 · submitted 2025-02-20 · 💻 cs.CL

LLM-based User Profile Management for Recommender System

Seunghwan Bang , Hwanjun Song This is my paper

Pith reviewed 2026-05-23 02:43 UTC · model grok-4.3

classification 💻 cs.CL

keywords LLM-based recommendationuser profile managementreview extractionsequential recommendationzero-shot recommendationAmazon datasetsprofile updating

0 comments

The pith

LLM extracts and updates user profiles from reviews to improve zero-shot recommendations.

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

The paper presents PURE as a framework that uses LLMs to extract preferences and features from user reviews, then maintains evolving profiles for recommendation. It targets the gap where most LLM recommenders rely only on purchase histories and struggle with long-term data due to token limits. The approach introduces a continuous sequential task where profiles update incrementally as new reviews arrive. Experiments on Amazon datasets show outperformance over prior LLM methods by better incorporating textual user information. A reader would care because this turns unstructured reviews into actionable, updatable profiles without model retraining.

Core claim

PURE consists of three components: a Review Extractor that pulls user preferences and product features from reviews, a Profile Updater that refines and summarizes profiles over time, and a Recommender that generates suggestions from the current profile. By maintaining these profiles across sequential review additions, the system leverages long-term user data while respecting token constraints, resulting in higher performance than existing LLM-based recommenders on Amazon datasets.

What carries the argument

The Profile Updater, which refines and summarizes extracted review information into evolving user profiles for ongoing recommendation use.

If this is right

Enables continuous, incremental profile updates as new reviews arrive over time.
Outperforms existing LLM-based methods on Amazon datasets in sequential recommendation.
Leverages long-term user information from textual reviews while managing token limits.
Supports zero-shot personalized recommendations without conventional training.

Where Pith is reading between the lines

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

The same extraction and updating process could apply to other text-heavy domains such as movie or news recommendation.
Hybrid systems might combine these LLM profiles with traditional collaborative filtering for further gains.
Testing on review histories longer than those in Amazon datasets would check scalability under heavier token pressure.

Load-bearing premise

The LLM components can accurately extract preferences and features from reviews without introducing substantial errors, hallucinations, or biases.

What would settle it

Running the full PURE pipeline on the Amazon datasets and finding that its recommendation metrics are no higher than those from purchase-history-only LLM baselines would falsify the claimed benefit of review-based profile management.

Figures

Figures reproduced from arXiv: 2502.14541 by Hwanjun Song, Seunghwan Bang.

**Figure 1.** Figure 1: Overall system of PURE. PURE incorporates reviews, ratings, and item interactions, whereas LLM Recommender handles only item interactions. By using the "Review Extractor" to identify key information and the "Profile Updater" to refine the user profile, PURE addresses scalability issue (ie growth of input token size). propose PURE, a novel LLM-based Profile Update for REcommender that constructs a user prof… view at source ↗

read the original abstract

The rapid advancement of Large Language Models (LLMs) has opened new opportunities in recommender systems by enabling zero-shot recommendation without conventional training. Despite their potential, most existing works rely solely on users' purchase histories, leaving significant room for improvement by incorporating user-generated textual data, such as reviews and product descriptions. Addressing this gap, we propose PURE, a novel LLM-based recommendation framework that builds and maintains evolving user profiles by systematically extracting and summarizing key information from user reviews. PURE consists of three core components: a Review Extractor for identifying user preferences and key product features, a Profile Updater for refining and updating user profiles, and a Recommender for generating personalized recommendations using the most current profile. To evaluate PURE, we introduce a continuous sequential recommendation task that reflects real-world scenarios by adding reviews over time and updating predictions incrementally. Our experimental results on Amazon datasets demonstrate that PURE outperforms existing LLM-based methods, effectively leveraging long-term user information while managing token limitations.

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.

Desk Editor's Note private letter to a colleague

PURE outlines a three-part pipeline to maintain review-based user profiles in LLM recommenders, but the abstract gives no metrics or details to support the outperformance claim.

read the letter

The paper's core contribution is PURE, a framework with a Review Extractor, Profile Updater, and Recommender that pulls preferences and features from user reviews to keep profiles current for sequential recommendations. This goes beyond simple history prompting by trying to manage long-term info without hitting token limits. What it does well is identify the gap in using review text and define a task where reviews arrive over time and predictions update incrementally. That setup feels closer to actual e-commerce use than static history-based tests. The main issue is the evidence. The abstract claims better results on Amazon datasets than other LLM methods, but it doesn't report any numbers, baselines, dataset sizes, or how they measured success. Without those, it's impossible to judge if the approach works or if the extractor is introducing errors. The assumption that the LLM can reliably summarize reviews without bias or hallucination is left untested in what's visible. This kind of work is aimed at researchers in recommender systems who are experimenting with LLMs. A reader looking for new prompting strategies or profile maintenance ideas might find the high-level design useful, but anyone wanting reproducible results will need the full experimental section. It deserves a serious referee because the idea is practical and the task definition is reasonable, even though the current description is light on data. If the experiments are sound, it could be a useful incremental step. Recommendation: If the full paper has proper ablations and error analysis on the extraction step, send it out for review. Otherwise it risks being too preliminary.

Referee Report

2 major / 1 minor

Summary. The paper proposes PURE, an LLM-based recommender framework with three components (Review Extractor to identify preferences and product features from reviews, Profile Updater to refine evolving profiles, and Recommender to generate recommendations from the current profile). It introduces a continuous sequential recommendation task on Amazon datasets that incrementally adds reviews and updates predictions, claiming that PURE outperforms prior LLM-based methods by leveraging long-term user information while managing token limits.

Significance. If the outperformance claim holds with rigorous validation, the work would be significant for zero-shot LLM recommenders by demonstrating a practical mechanism to incorporate and maintain user-generated textual data over time, addressing a gap in methods that rely only on purchase histories.

major comments (2)

[Abstract] Abstract: the central claim that 'our experimental results on Amazon datasets demonstrate that PURE outperforms existing LLM-based methods' supplies no metrics, baselines, statistical tests, dataset sizes, or implementation details, leaving the empirical result unsupported by visible evidence.
[Abstract] Abstract (and implied experimental section): no quantitative validation such as human agreement rates, error analysis on extracted attributes, or ablation studies removing the LLM extraction components is reported, so the assumption that the Review Extractor and Profile Updater produce faithful summaries without substantial hallucinations or biases remains untested and load-bearing for attributing any gains to the framework.

minor comments (1)

The description of the continuous sequential recommendation task would benefit from an explicit definition of the evaluation protocol (e.g., how predictions are updated and what constitutes a 'review added over time').

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback highlighting areas where the abstract and experimental validation can be strengthened. We address each major comment below and will revise the manuscript accordingly.

read point-by-point responses

Referee: [Abstract] Abstract: the central claim that 'our experimental results on Amazon datasets demonstrate that PURE outperforms existing LLM-based methods' supplies no metrics, baselines, statistical tests, dataset sizes, or implementation details, leaving the empirical result unsupported by visible evidence.

Authors: We agree that the abstract should be more self-contained and provide concrete support for the performance claim. In the revised version, we will update the abstract to include specific metrics (such as relative improvements over baselines), the Amazon datasets and their sizes, the main baselines, and a note on statistical testing. Full experimental details remain in Section 4, but the abstract will now summarize key quantitative evidence. revision: yes
Referee: [Abstract] Abstract (and implied experimental section): no quantitative validation such as human agreement rates, error analysis on extracted attributes, or ablation studies removing the LLM extraction components is reported, so the assumption that the Review Extractor and Profile Updater produce faithful summaries without substantial hallucinations or biases remains untested and load-bearing for attributing any gains to the framework.

Authors: We acknowledge that the current manuscript does not include direct quantitative validation of the extraction components via human agreement rates, error analysis, or component ablations. While the overall recommendation gains provide supporting evidence, we agree these additions would strengthen attribution of results to the framework. We will incorporate ablation studies (with and without the Review Extractor and Profile Updater), a sample-based error analysis on extracted attributes, and human agreement rates from a targeted annotation in the revised manuscript. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical framework evaluated against external baselines

full rationale

The paper describes an LLM-based recommender framework (PURE) with three components and reports experimental outperformance on Amazon datasets in a sequential recommendation task. No equations, parameter fits, or derivation chains appear in the provided text. Claims rest on comparisons to prior external methods rather than any self-referential construction, fitted inputs renamed as predictions, or load-bearing self-citations. This matches the default case of a self-contained empirical paper with independent external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract supplies no explicit free parameters, mathematical axioms, or new physical entities; the contribution is a procedural framework whose internal LLM calls are treated as black boxes.

pith-pipeline@v0.9.0 · 5692 in / 922 out tokens · 39285 ms · 2026-05-23T02:43:10.456881+00:00 · methodology

discussion (0)

Forward citations

Cited by 2 Pith papers

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

Bridging Textual Profiles and Latent User Embeddings for Personalization
cs.IR 2026-05 unverdicted novelty 6.0

BLUE aligns LLM-generated textual user profiles with embedding-based recommendation objectives via reinforcement learning and next-item text supervision, yielding better zero-shot performance and cross-domain transfer...
Democratizing Large-Scale Re-Optimization with LLM-Guided Model Patches
cs.AI 2026-05 unverdicted novelty 5.0

An LLM agent converts user prompts into optimization-model patches and selects primal-based re-optimization methods from a toolbox to produce feasible solutions for dynamic supply-chain and exam-scheduling problems.

Reference graph

Works this paper leans on

30 extracted references · 30 canonical work pages · cited by 2 Pith papers · 5 internal anchors

[1]

LLaMA: Open and Efficient Foundation Language Models

H. Touvron, T. Lavril, et al., Llama: Open and efficient foundation language models, ArXiv:2302.13971 (2023)

work page internal anchor Pith review Pith/arXiv arXiv 2023
[2]

The Llama 3 Herd of Models

A. Dubey, A. Jauhri, A. Pandey, A. Kadian, A. Al-Dahle, A. Letman, A. Mathur, A. Schelten, A. Yang, A. Fan, et al., The llama 3 herd of models, ArXiv:2407.21783 (2024)

work page internal anchor Pith review Pith/arXiv arXiv 2024
[3]

GPT-4 Technical Report

J. Achiam, S. Adler, S. Agarwal, L. Ahmad, I. Akkaya, F. L. Aleman, D. Almeida, J. Altenschmidt, S. Altman, S. Anadkat, et al., Gpt-4 technical report, ArXiv:2303.08774 (2023)

work page internal anchor Pith review Pith/arXiv arXiv 2023
[4]

G. Team, M. Riviere, S. Pathak, P. G. Sessa, C. Hardin, S. Bhupatiraju, L. Hussenot, T. Mesnard, B. Shahriari, A. Ramé, et al., Gemma 2: Improving open language models at a practical size, ArXiv:2408.00118 (2024)

work page internal anchor Pith review Pith/arXiv arXiv 2024
[5]

Lewis, Y

M. Lewis, Y. Liu, et al., Bart: Denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension, in: Proceedings of the Annual Meeting of the Association for Computational Linguistics (ACL), 2020

work page 2020
[6]

Karpukhin, B

V. Karpukhin, B. Oguz, S. Min, P. Lewis, L. Wu, S. Edunov, D. Chen, W.-t. Yih, Dense passage retrieval for open-domain question answering, in: Proceedings of Empirical Methods in Natural Language Processing (EMNLP), 2020, pp. 6769–6781

work page 2020
[7]

T. B. Brown, B. Mann, N. Ryder, et al., Language models are few-shot learners, in: Proceedings of Advances in Neural Information Processing Systems (NeurIPS), volume 33, 2020, pp. 1877–1901

work page 2020
[8]

Lewis, E

P. Lewis, E. Perez, A. Piktus, F. Petroni, V. Karpukhin, N. Goyal, H. Küttler, M. Lewis, W.-t. Yih, T. Rocktäschel, et al., Retrieval-augmented generation for knowledge-intensive nlp tasks, in: Proceedings of Advances in Neural Information Processing Systems (NeurIPS), volume 33, 2020, pp. 9459–9474

work page 2020
[9]

X. He, L. Liao, H. Zhang, L. Nie, X. Hu, T.-S. Chua, Neural collaborative filtering, in: Proceedings of International Conference on World Wide Web (WWW), 2017, pp. 173–182

work page 2017
[10]

W.-C. Kang, J. McAuley, Self-attentive sequential recommendation, in: Proceedings of International Conference on Data Mining (ICDM), IEEE, 2018, pp. 197–206

work page 2018
[11]

X. He, K. Deng, X. Wang, Y. Li, Y. Zhang, M. Wang, Lightgcn: Simplifying and powering graph convolution network for recommendation, in: Proceedings of international ACM SIGIR conference on research and development in Information Retrieval, 2020, pp. 639–648

work page 2020
[12]

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

work page 2024
[13]

W. Wei, X. Ren, J. Tang, Q. Wang, L. Su, S. Cheng, J. Wang, D. Yin, C. Huang, Llmrec: Large language models with graph augmentation for recommendation, in: Proceedings of the ACM International Conference on Web Search and Data Mining, 2024, pp. 806–815

work page 2024
[14]

X. Ren, W. Wei, L. Xia, L. Su, S. Cheng, J. Wang, D. Yin, C. Huang, Representation learning with large language models for recommendation, in: Proceedings of the ACM on Web Conference, 2024, pp. 3464–3475

work page 2024
[15]

Z. He, Z. Xie, R. Jha, H. Steck, D. Liang, Y. Feng, B. P. Majumder, N. Kallus, J. McAuley, Large lan- guage models as zero-shot conversational recommenders, in: Proceedings of the ACM International Conference on Information and Knowledge Management (CIKM), 2023, pp. 720–730

work page 2023
[16]

J. Zhai, X. Zheng, C.-D. Wang, H. Li, Y. Tian, Knowledge prompt-tuning for sequential recommen- dation, in: Proceedings of ACM International Conference on Multimedia, 2023, pp. 6451–6461

work page 2023
[17]

N. F. Liu, K. Lin, J. Hewitt, A. Paranjape, M. Bevilacqua, F. Petroni, P. Liang, Lost in the middle: How language models use long contexts, Transactions of the Association for Computational Linguistics 12 (2024) 157–173

work page 2024
[18]

X. Li, S. Wang, S. Zeng, Y. Wu, Y. Yang, A survey on llm-based multi-agent systems: workflow, infrastructure, and challenges, Vicinagearth 1 (2024) 9

work page 2024
[19]

Y. Ding, L. L. Zhang, C. Zhang, Y. Xu, N. Shang, J. Xu, F. Yang, M. Yang, Longrope: Extending llm context window beyond 2 million tokens, ArXiv:2402.13753 (2024)

work page internal anchor Pith review Pith/arXiv arXiv 2024
[20]

Zhong, L

W. Zhong, L. Guo, Q. Gao, H. Ye, Y. Wang, Memorybank: Enhancing large language models with long-term memory, in: Proceedings of the Association for the Advancement of Artificial Intelligence (AAAI), volume 38, 2024, pp. 19724–19731

work page 2024
[21]

J. Ni, J. Li, J. McAuley, Justifying recommendations using distantly-labeled reviews and fine- grained aspects, in: Proceedings of the 2019 conference on empirical methods in natural language processing and the 9th international joint conference on natural language processing (EMNLP- IJCNLP), 2019, pp. 188–197

work page 2019
[22]

K. Bao, J. Zhang, Y. Zhang, W. Wang, F. Feng, X. He, Tallrec: An effective and efficient tuning framework to align large language model with recommendation, in: Proceedings of the 17th ACM Conference on Recommender Systems, 2023, pp. 1007–1014

work page 2023
[23]

S. Wang, L. Hu, Y. Wang, L. Cao, Q. Z. Sheng, M. Orgun, Sequential recommender systems: Challenges, progress and prospects, in: Proceedings of International Joint Conference on Artificial Intelligence Organization (IJCAI), 2019, pp. 6332–6338

work page 2019
[24]

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

work page 2019
[25]

Hidasi, A

B. Hidasi, A. Karatzoglou, Recurrent neural networks with top-k gains for session-based rec- ommendations, in: Proceedings of the 27th ACM international conference on information and knowledge management, 2018, pp. 843–852

work page 2018
[26]

S. Kim, H. Kang, S. Choi, D. Kim, M. Yang, C. Park, Large language models meet collaborative filtering: An efficient all-round llm-based recommender system, in: Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2024, pp. 1395–1406

work page 2024
[27]

J. Kim, H. Kim, H. Cho, S. Kang, B. Chang, J. Yeo, D. Lee, driven personalized preference reasoning with large language models for recommendation, arXiv preprint arXiv:2408.06276 (2024)

work page arXiv 2024
[28]

Wang, E.-P

L. Wang, E.-P. Lim, Zero-shot next-item recommendation using large pretrained language models, ArXiv:2304.03153 (2023)

work page arXiv 2023
[29]

S. Dai, N. Shao, H. Zhao, W. Yu, Z. Si, C. Xu, Z. Sun, X. Zhang, J. Xu, Uncovering chatgpt’s capabil- ities in recommender systems, in: Proceedings of the 17th ACM Conference on Recommender Systems, 2023, pp. 1126–1132

work page 2023
[30]

Zhang, R

J. Zhang, R. Xie, Y. Hou, X. Zhao, L. Lin, J.-R. Wen, Recommendation as instruction following: A large language model empowered recommendation approach, ACM Transactions on Information Systems (2023)

work page 2023

[1] [1]

LLaMA: Open and Efficient Foundation Language Models

H. Touvron, T. Lavril, et al., Llama: Open and efficient foundation language models, ArXiv:2302.13971 (2023)

work page internal anchor Pith review Pith/arXiv arXiv 2023

[2] [2]

The Llama 3 Herd of Models

A. Dubey, A. Jauhri, A. Pandey, A. Kadian, A. Al-Dahle, A. Letman, A. Mathur, A. Schelten, A. Yang, A. Fan, et al., The llama 3 herd of models, ArXiv:2407.21783 (2024)

work page internal anchor Pith review Pith/arXiv arXiv 2024

[3] [3]

GPT-4 Technical Report

J. Achiam, S. Adler, S. Agarwal, L. Ahmad, I. Akkaya, F. L. Aleman, D. Almeida, J. Altenschmidt, S. Altman, S. Anadkat, et al., Gpt-4 technical report, ArXiv:2303.08774 (2023)

work page internal anchor Pith review Pith/arXiv arXiv 2023

[4] [4]

G. Team, M. Riviere, S. Pathak, P. G. Sessa, C. Hardin, S. Bhupatiraju, L. Hussenot, T. Mesnard, B. Shahriari, A. Ramé, et al., Gemma 2: Improving open language models at a practical size, ArXiv:2408.00118 (2024)

work page internal anchor Pith review Pith/arXiv arXiv 2024

[5] [5]

Lewis, Y

M. Lewis, Y. Liu, et al., Bart: Denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension, in: Proceedings of the Annual Meeting of the Association for Computational Linguistics (ACL), 2020

work page 2020

[6] [6]

Karpukhin, B

V. Karpukhin, B. Oguz, S. Min, P. Lewis, L. Wu, S. Edunov, D. Chen, W.-t. Yih, Dense passage retrieval for open-domain question answering, in: Proceedings of Empirical Methods in Natural Language Processing (EMNLP), 2020, pp. 6769–6781

work page 2020

[7] [7]

T. B. Brown, B. Mann, N. Ryder, et al., Language models are few-shot learners, in: Proceedings of Advances in Neural Information Processing Systems (NeurIPS), volume 33, 2020, pp. 1877–1901

work page 2020

[8] [8]

Lewis, E

P. Lewis, E. Perez, A. Piktus, F. Petroni, V. Karpukhin, N. Goyal, H. Küttler, M. Lewis, W.-t. Yih, T. Rocktäschel, et al., Retrieval-augmented generation for knowledge-intensive nlp tasks, in: Proceedings of Advances in Neural Information Processing Systems (NeurIPS), volume 33, 2020, pp. 9459–9474

work page 2020

[9] [9]

X. He, L. Liao, H. Zhang, L. Nie, X. Hu, T.-S. Chua, Neural collaborative filtering, in: Proceedings of International Conference on World Wide Web (WWW), 2017, pp. 173–182

work page 2017

[10] [10]

W.-C. Kang, J. McAuley, Self-attentive sequential recommendation, in: Proceedings of International Conference on Data Mining (ICDM), IEEE, 2018, pp. 197–206

work page 2018

[11] [11]

X. He, K. Deng, X. Wang, Y. Li, Y. Zhang, M. Wang, Lightgcn: Simplifying and powering graph convolution network for recommendation, in: Proceedings of international ACM SIGIR conference on research and development in Information Retrieval, 2020, pp. 639–648

work page 2020

[12] [12]

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

work page 2024

[13] [13]

W. Wei, X. Ren, J. Tang, Q. Wang, L. Su, S. Cheng, J. Wang, D. Yin, C. Huang, Llmrec: Large language models with graph augmentation for recommendation, in: Proceedings of the ACM International Conference on Web Search and Data Mining, 2024, pp. 806–815

work page 2024

[14] [14]

X. Ren, W. Wei, L. Xia, L. Su, S. Cheng, J. Wang, D. Yin, C. Huang, Representation learning with large language models for recommendation, in: Proceedings of the ACM on Web Conference, 2024, pp. 3464–3475

work page 2024

[15] [15]

Z. He, Z. Xie, R. Jha, H. Steck, D. Liang, Y. Feng, B. P. Majumder, N. Kallus, J. McAuley, Large lan- guage models as zero-shot conversational recommenders, in: Proceedings of the ACM International Conference on Information and Knowledge Management (CIKM), 2023, pp. 720–730

work page 2023

[16] [16]

J. Zhai, X. Zheng, C.-D. Wang, H. Li, Y. Tian, Knowledge prompt-tuning for sequential recommen- dation, in: Proceedings of ACM International Conference on Multimedia, 2023, pp. 6451–6461

work page 2023

[17] [17]

N. F. Liu, K. Lin, J. Hewitt, A. Paranjape, M. Bevilacqua, F. Petroni, P. Liang, Lost in the middle: How language models use long contexts, Transactions of the Association for Computational Linguistics 12 (2024) 157–173

work page 2024

[18] [18]

X. Li, S. Wang, S. Zeng, Y. Wu, Y. Yang, A survey on llm-based multi-agent systems: workflow, infrastructure, and challenges, Vicinagearth 1 (2024) 9

work page 2024

[19] [19]

Y. Ding, L. L. Zhang, C. Zhang, Y. Xu, N. Shang, J. Xu, F. Yang, M. Yang, Longrope: Extending llm context window beyond 2 million tokens, ArXiv:2402.13753 (2024)

work page internal anchor Pith review Pith/arXiv arXiv 2024

[20] [20]

Zhong, L

W. Zhong, L. Guo, Q. Gao, H. Ye, Y. Wang, Memorybank: Enhancing large language models with long-term memory, in: Proceedings of the Association for the Advancement of Artificial Intelligence (AAAI), volume 38, 2024, pp. 19724–19731

work page 2024

[21] [21]

J. Ni, J. Li, J. McAuley, Justifying recommendations using distantly-labeled reviews and fine- grained aspects, in: Proceedings of the 2019 conference on empirical methods in natural language processing and the 9th international joint conference on natural language processing (EMNLP- IJCNLP), 2019, pp. 188–197

work page 2019

[22] [22]

K. Bao, J. Zhang, Y. Zhang, W. Wang, F. Feng, X. He, Tallrec: An effective and efficient tuning framework to align large language model with recommendation, in: Proceedings of the 17th ACM Conference on Recommender Systems, 2023, pp. 1007–1014

work page 2023

[23] [23]

S. Wang, L. Hu, Y. Wang, L. Cao, Q. Z. Sheng, M. Orgun, Sequential recommender systems: Challenges, progress and prospects, in: Proceedings of International Joint Conference on Artificial Intelligence Organization (IJCAI), 2019, pp. 6332–6338

work page 2019

[24] [24]

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

work page 2019

[25] [25]

Hidasi, A

B. Hidasi, A. Karatzoglou, Recurrent neural networks with top-k gains for session-based rec- ommendations, in: Proceedings of the 27th ACM international conference on information and knowledge management, 2018, pp. 843–852

work page 2018

[26] [26]

S. Kim, H. Kang, S. Choi, D. Kim, M. Yang, C. Park, Large language models meet collaborative filtering: An efficient all-round llm-based recommender system, in: Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 2024, pp. 1395–1406

work page 2024

[27] [27]

J. Kim, H. Kim, H. Cho, S. Kang, B. Chang, J. Yeo, D. Lee, driven personalized preference reasoning with large language models for recommendation, arXiv preprint arXiv:2408.06276 (2024)

work page arXiv 2024

[28] [28]

Wang, E.-P

L. Wang, E.-P. Lim, Zero-shot next-item recommendation using large pretrained language models, ArXiv:2304.03153 (2023)

work page arXiv 2023

[29] [29]

S. Dai, N. Shao, H. Zhao, W. Yu, Z. Si, C. Xu, Z. Sun, X. Zhang, J. Xu, Uncovering chatgpt’s capabil- ities in recommender systems, in: Proceedings of the 17th ACM Conference on Recommender Systems, 2023, pp. 1126–1132

work page 2023

[30] [30]

Zhang, R

J. Zhang, R. Xie, Y. Hou, X. Zhao, L. Lin, J.-R. Wen, Recommendation as instruction following: A large language model empowered recommendation approach, ACM Transactions on Information Systems (2023)

work page 2023