Pith. sign in

REVIEW 4 major objections 8 minor 3 cited by

RecLM: Recommendation Instruction Tuning

T0 review · 4 major / 8 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read RecLM claims that instruction-tuning a language model to write user and item profiles, then adding those profiles to existing collaborative-filtering recommenders, sharply improves ranking accuracy and cold-start generalization.

desk verdict RecLM is a serious, mostly solid instruction-tuning framework for LLM-generated user/item profiles, with real gains and one reward-model concern that needs a validation check. read the letter →

arxiv 2412.19302 v3 pith:TK3Y3E6J submitted 2024-12-26 cs.IR

Yangqin Jiang , Yuhao Yang , Lianghao Xia , Da Luo , Kangyi Lin , Chao Huang This is my paper · ORCID
classification cs.IR
keywords recommendationinstructiontuningcollaborativefilteringlargelanguagemodelscold-startzero-shotgeneralizationreinforcementlearninguserprofilingitem
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

RecLM is a model-agnostic way to give collaborative-filtering recommenders a language-model-assisted path into cold-start scenarios. The paper claims that instruction-tuning an open LLM to write structured user profiles (three identities and three interests) and item profiles (five identities and five interests), then adding those profiles as extra input features to existing ID-based recommenders, reliably beats the same recommenders using only raw text and IDs. The gains are largest in zero-shot cold-start evaluation: on Netflix, SGL's Recall@20 rises from 0.0385 to 0.1126, and on MIND, LightGCN's full-shot Recall@20 roughly doubles from 0.0389 to 0.0788. The paper attributes these improvements to injecting collaborative neighborhood information into the profile text, followed by a reinforcement-learning stage that reduces noise and over-smoothing. If true, this gives a practical plug-in for sparse and cold-start recommendation.

What carries the argument

The load-bearing mechanism is the collaborative profile-generation pipeline: a two-turn dialogue that injects higher-order user-user and user-item relations into LLM prompts, a mask-based multi-turn tuning objective so both the profile response and the Yes/No interaction response contribute loss, and a reward-model-plus-PPO stage that refines the generated profiles. The profile format is fixed—three identities and three interests for users, five identities and five interests for items—and is produced by Llama2-7B fine-tuned with LoRA. This pipeline converts raw item text into the semantic features that the fusion MLP in Eq. (2) combines with collaborative embeddings.

What would settle it

Run the zero-shot Netflix SGL setup again but replace every RecLM-generated user profile with a fixed generic template while keeping all other components identical; if Recall@20 stays close to 0.1126, then the claimed personalization of profiles is not what drives the gain and the central claim would collapse, whereas a fall back toward the 0.0385 baseline would confirm that the profile content is doing the work.

Watch

Extended reading notes

Core claim

The paper's central claim is that a language model can be made to produce recommendation-grade profiles by tuning it on collaborative relationships, not just on item text. The two-turn instruction tuning first asks the LLM to profile a target user together with similar users, then asks it to predict a held-out interaction using those profiles; masking lets both response turns contribute to weight updates. A reward model trained on ChatGPT-generated positives and hand-constructed negatives (missing information, duplicated content, substituted profiles of similar users) then drives PPO refinement. The resulting profiles are fused into base recommenders through a dual-MLP projection, replacing pure-ID user embeddings with text-plus-profile representations. Across full-shot, zero-shot, and industrial settings, integrating RecLM improves every tested backbone over its base variant, with most differences marked statistically significant at $p<0.05$.

Load-bearing premise

The load-bearing premise is that the language-model-written profiles—three identities and three interests per user, five per item—are faithful enough to represent genuine preferences, and that the reward model used to refine them rewards true personalization rather than generic text that merely matches its training patterns.

Editorial extensions

If this is right

  • Any ID-based recommender with access to item text can be upgraded by appending RecLM-generated profiles; no retraining of the backbone architecture is required.
  • In fully cold-start settings, unseen items can be scored through their text-derived representations, which explains why zero-shot gains are large even for graph recommenders that normally cannot embed new IDs.
  • Because profiles are reusable offline features, online systems can refresh them daily or weekly at modest cost; per-epoch training overhead stays under about 25% and below 10% for most GNN backbones.
  • The reinforcement-learning refinement step is necessary for the result: without it, profiles absorb too much neighbor information and performance drops in both the MIND and Netflix ablations.
  • Consistent improvements across matrix factorization, neural, graph, and contrastive backbones support the claim that the approach is model-agnostic rather than tied to one recommender architecture.

Reading between the lines

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

  • One implied test is to vary the number of similar users shown in the prompt; if too many neighbors degrade profiles even with RL, the robustness boundary is the neighborhood size and similarity metric rather than the LLM itself.
  • The reward model is trained on synthetic negatives, so a natural extension is to replace them with human preference judgments; the performance gap between synthetic and human rewards would indicate how much of the RL gain reflects genuine personalization.
  • The same instruction-tuning recipe could transfer to multimodal item side information, such as images, audio, or video, though the paper only reports text-side experiments.
  • Because the profiles are written in a fixed text format and fused externally, a single tuned LLM might serve as a cross-dataset profile bank; the paper evaluates per-dataset, so cross-dataset reuse remains an open question.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 8 minor

Summary. The paper proposes RecLM, a model-agnostic instruction-tuning framework that uses an LLM to generate user and item profiles from textual side information and collaborative neighborhoods, then refines profile generation with a reinforcement-learning stage. The generated profiles are fused into existing collaborative-filtering recommenders through MLP projection layers. The authors evaluate RecLM on MIND, Netflix, and a proprietary Industrial dataset, with five backbone recommenders (BiasMF, NCF, LightGCN, SGL, SimGCL), in both full-shot and zero-shot settings, and report large relative improvements in Recall@20/40 and NDCG@20/40, along with ablations of the instruction-tuning components, an efficiency study, and a comparison with LLMRec. The code is publicly available.

Significance. If the results hold, RecLM offers a practical recipe for injecting LLM-generated profile signals into existing CF recommenders without retraining the backbone, with a plausible mechanism for improving cold-start performance. The paper's strengths include the breadth of the evaluation (three datasets, five backbones, two settings), the release of code, and the ablation design that separates knowledge-distillation, collaborative instruction tuning, and RL-based refinement. The central risk is that the RL contribution rests on a reward model whose validity is not demonstrated, and that the headline claim of consistent improvement is contradicted by several entries in the main table. With additional validation, this could be a useful contribution; in its current form the evidence is incomplete.

major comments (4)
  1. [Section 3.2, Table 1] The claim that integrating RecLM 'consistently' leads to enhanced performance is contradicted by entries in the same table. In the full-shot MIND block, BiasMF N@20 drops from 0.0311 to 0.0272 (Improve: 12.54% down) and NCF N@20/N@40 drop from 0.0325/0.0445 to 0.0288/0.0414; in the zero-shot Industrial block, SimGCL R@40 decreases by 2.14% and N@40 is flat. The Improve row reports these negative changes without discussion, while the prose in observations (i)-(iii) asserts consistent gains. The authors should either explain these specific cases (e.g., metric variance, profile noise, or a trade-off) or revise the claim to a majority-of-cases statement supported by a formal significance analysis.
  2. [Section 2.4, Eq. (6), Appendix 6.6, Fig. 9] The reward model r_theta is trained only on ChatGPT-generated positive profiles and hand-crafted negative profiles (missing information, duplicated content, substituted similar-user profiles). No evidence is provided that r_theta's scores correlate with human judgments of profile quality or with downstream recommendation performance on held-out data. Since the PPO objective in Eq. (8) maximizes r_theta minus a KL penalty, the RL stage may simply learn to emit outputs that resemble the ChatGPT training distribution or conform to the three-identity/three-interest template, rather than more personalized profiles. This is load-bearing because RQ2 and RQ6 attribute part of the gain specifically to RL. Please add (i) reward-model accuracy or human-correlation results, (ii) an ablation replacing the learned reward with a simple rule-based or random reward, and (iii) a quantitative check that RL-produced profiles are more personalized rather than more template-conforming; the current case study in Fig. 4 is a single example and does not resolve this.
  3. [Section 2.2, Eq. (2), Appendix 6.6, Figs. 7-8] The central mechanism is the fixed-format profile (3 identities plus 3 interests for users, 5 plus 5 for items), but the paper does not validate that this representation is sufficient to convey preference structure. The only evidence for the profile's informativeness is downstream recommendation performance and one qualitative case study. Please report profile-level quality metrics, an ablation varying profile lengths, or a comparison against profiles generated without the fixed-format constraint. Without such evidence, the attribution of the gains to the profile content, rather than to the additional text features alone, is not fully established.
  4. [Sections 3.2-3.4] Statistical significance is reported only as p<0.05, with no specification of the test used, the number of random seeds, or the variance across runs. Table 2 and Figures 2-3 do not report error bars or significance levels. Given that some entries in Table 1 show decreases and others show very large relative improvements, the robustness of the central claim cannot be assessed. Please report means and standard deviations over at least three seeds, and state the exact statistical test used for the significance markers.
minor comments (8)
  1. [Section 2.3.1] There is a typo in the sentence describing ChatGPT-generated profiles: 'interaction patters among suers and items' should read 'users and items.'
  2. [Appendix 6.6, Fig. 8] The item-profile generation instruction says 'please generate the user profile of this target item'; this should be 'item profile' to avoid confusion between user profiles and item profiles.
  3. [Table 1] In the NCF full-shot Industrial row, the N@40 Augment value is shown as '0.0108∗ ∗' with an extra asterisk; please clean up the notation and use a consistent convention for negative improvements.
  4. [Figures 5-8] Several instruction-design figures appear to contain duplicated or repeated panels (especially Figures 5 and 6, and Figures 7 and 8); please replace them with clean, non-duplicated diagrams.
  5. [Section 3.5] The efficiency analysis reports only per-epoch training time of the backbone recommender with and without RecLM. The offline cost of LLM profile generation (ChatGPT API calls, LoRA fine-tuning, and PPO training) is not reported, which makes the practicality claim in RQ4 incomplete.
  6. [Section 6.5] Implementation details for the LLM tuning pipeline are missing: LoRA rank and alpha, LLM learning rate and epochs, reward model architecture, PPO hyperparameters (beta, clip ratio, number of epochs), and total GPU hours.
  7. [Section 3.6] The comparison with existing LLM-enhanced recommenders is limited to LLMRec; the related work discusses RLMRec, InstructRec, and TALLRec, but these are not included in the experiments. A broader comparison or an explicit explanation for their exclusion would strengthen RQ5.
  8. [Appendix 6.1] The Industrial dataset is described only as anonymized; beyond the statistics in Table 4, there is no information about item types, interaction distribution, or how the proprietary data were sampled, which limits reproducibility.

Circularity Check

1 steps flagged · score 4.0 of 10

RL reward model is fitted to ChatGPT-generated positives, so the claimed RL-driven profile-quality gains are partly self-referential; the main recommendation results remain externally benchmarked.

  1. fitted input called prediction [Sec. 2.4 (Eq. 6, Eq. 8; Reward Model and PPO); Appendix 6.6 (Reward Model Training, Fig. 9)]
    "We obtain R+ through ChatGPT and create R− through two strategies: Diverse Negative Sampling ... Profile Substitution ... The final reward function ... R(Ri|Qi) = ˆr(Ri|Qi) − βDKL(Mθ(Qi)||M0(Qi)) (8)"

    The reward model rθ is fitted by Eq. 6 using R+ drawn from ChatGPT and R− constructed by the authors (missing information, duplication, substitution with similar users' profiles). PPO then maximizes Eq. 8, in which the same rθ is the only quality signal. Therefore 'high-quality/personalized profile' is, by construction, a profile that the fitted reward model ranks above hand-crafted negatives, i.e., one resembling the ChatGPT-derived distribution that also produced the instruction-tuning targets (Sec. 2.3.1).

full rationale

The central recommendation-performance claim is not circular: the Base-versus-Augment comparison is an external evaluation, and the full/zero-shot metrics are computed on held-out test interactions. The two-turn instruction tuning uses ground-truth interaction labels from the training data, not the predicted quantities, so no fitted-input-called-prediction reduction applies to the main results. No load-bearing self-citation chain appears; comparisons to LLMRec and other baselines are external, and the method does not invoke a uniqueness theorem from the authors' prior work. The one structurally circular element is the RL reward model: its positive labels are ChatGPT outputs, which come from the same LLM distribution that generated the supervised instruction-tuning data, so the RL objective's notion of profile quality is self-defined. This weakens the internal claim that RL improves personalization as such, but it does not by itself force the downstream benchmark improvements, leaving the overall circularity partial rather than total.

Assumptions & free parameters 3 free parameters · 3 assumptions · 1 invented entities

The central claim does not rest on a derived mathematical result. It rests on measured performance improvements from an LLM-based profiling pipeline. The free parameters are mostly architectural and prompt choices that are not fully specified or ablated. The most load-bearing assumptions are the sufficiency of text features, the validity of LightGCN-based neighborhoods, and the fixed profile format. The reward model introduces a mild circularity because its training distribution is the same ChatGPT-generated profile distribution used to generate positives.

free parameters (3)
  • number of similar users k = not reported
    The two-turn instruction construction and inference prompts use a set of similar users {un}; the paper does not report how many neighbors are used or how this number was chosen.
  • item profile length (5 identities and 5 interests) = 5
    The item profile generation prompt hard-codes 5 identities and 5 interests, while user profiles use 3; this choice is not justified or ablated.
  • reward model capacity and training set size = not reported
    The reward model architecture, parameter count, and number of positive/negative pairs are not specified, and these choices affect the RL refinement behavior.
assumptions (3)
  • domain assumption Similar users identified by LightGCN embeddings are reliable sources of collaborative signal for profile generation.
    The two-turn instruction tuning and inference prompts construct collaborative neighborhoods from LightGCN user embeddings without validating the quality of those neighborhoods (Sec. 2.3.2).
  • domain assumption Textual descriptions of items are sufficient side information to represent user preferences and item semantics.
    The whole framework replaces ID-only embeddings with text features and profiles; this is a stated design choice, but its validity is only tested indirectly through final performance (Sec. 2.2).
  • ad hoc to paper The profile format (3 identities and 3 interests for users, 5 for items) is lossless enough to convey preference structure.
    The restrictive output format is enforced in every prompt and no ablation studies how the fixed-length profile affects quality (Appendix 6.6).
invented entities (1)
  • LLM-generated user and item profiles
    purpose: The profiles act as auxiliary semantic features fused with ID embeddings to improve cold-start recommendation.
    The profile quality is only assessed through end-to-end recommendation metrics; there is no external falsifiable handle such as a directly measured profile accuracy.

how reviews work

0 comments
Cite this review

Pith. "Pith review of RecLM: Recommendation Instruction Tuning." pith.science (2026). https://pith.science/paper/TK3Y3E6J

@misc{pith2026241219302,
  author       = {Pith},
  title        = {Pith review of: RecLM: Recommendation Instruction Tuning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TK3Y3E6J}},
  note         = {Machine review of arXiv:2412.19302}
}
abstract

Modern recommender systems aim to deeply understand users' complex preferences through their past interactions. While deep collaborative filtering approaches using Graph Neural Networks (GNNs) excel at capturing user-item relationships, their effectiveness is limited when handling sparse data or zero-shot scenarios, primarily due to constraints in ID-based embedding functions. To address these challenges, we propose a model-agnostic recommendation instruction-tuning paradigm that seamlessly integrates large language models with collaborative filtering. Our proposed $\underline{Rec}$ommendation $\underline{L}$anguage $\underline{M}$odel (RecLM) enhances the capture of user preference diversity through a carefully designed reinforcement learning reward function that facilitates self-augmentation of language models. Comprehensive evaluations demonstrate significant advantages of our approach across various settings, and its plug-and-play compatibility with state-of-the-art recommender systems results in notable performance enhancements. The implementation of our RecLM framework is publicly available at: https://github.com/HKUDS/RecLM.

Figures

Figures reproduced from arXiv: 2412.19302 by the authors.

Figure 1
Figure 1. The Overall Framework of the Proposed RecLM. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Ablation study on the LLM tuning techniques in the RecLM framework. [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Performance Comparison with LLMRec. We further compare RecLM with the existing work LLMRec (Wei et al., 2024), which also enhances recommendation systems using LLMs, to highlight the superiority of our proposed instruction-tuning technique. The experimen￾tal results are presented in [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: Generated profiles w/ and w/o RL. To intuitively explore the contribution of rein￾forcement learning to the personalization of gen￾erated profiles, we conducted a case study using the MIND dataset. In this study, as shown in [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 6
Figure 6. Figure 6: Instruction designs for two-turn instruction tuning. [PITH_FULL_IMAGE:figures/full_fig_p015_6.png]
Figure 8
Figure 8. Figure 8: Instruction designs for item profile generation. [PITH_FULL_IMAGE:figures/full_fig_p016_8.png]
Figure 9
Figure 9. Figure 9: Positive/Negative responses construction for reward model training. [PITH_FULL_IMAGE:figures/full_fig_p016_9.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 3 Pith papers

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

  1. GARDRec: Decision-Level Graph Grounding for Large Language Model Recommendation

    cs.IR 2026-08 conditional novelty 6.0 of 10

    GARDRec improves LLM-based next-item ranking by grounding decisions in knowledge-graph embeddings, personalized graph contexts, and late-stage scoring rather than prompt text.

  2. RecGPT: A Foundation Model for Sequential Recommendation

    cs.IR 2025-06 conditional novelty 5.0 of 10

    RecGPT turns item descriptions into shared discrete tokens and trains an autoregressive transformer to predict the next item's tokens, enabling zero-shot recommendations in unseen domains.

  3. Large Language Model Enhanced Recommender Systems: A Survey

    cs.IR 2024-12 unverdicted novelty 4.0 of 10

    A survey organizing LLM-enhanced recommender systems into knowledge, interaction, and model enhancement, and tracing a shift from explicit text to implicit embeddings and fine-tuned open-source LLMs.

Reference graph

Works this paper leans on

39 extracted references · 32 canonical work pages · cited by 3 Pith papers

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    Tallrec: An effective and efficient tuning framework to align large language model with recommendation

    Keqin Bao, Jizhi Zhang, Yang Zhang, Wenjie Wang, Fuli Feng, and Xiangnan He. Tallrec: An effective and efficient tuning framework to align large language model with recommendation. In RecSys, pp.\ 1007--1014, 2023

    work page 2023

  3. [3]

    Revisiting graph based collaborative filtering: A linear residual graph convolutional network approach

    Lei Chen, Le Wu, Richang Hong, Kun Zhang, and Meng Wang. Revisiting graph based collaborative filtering: A linear residual graph convolutional network approach. In AAAI'20

    work page

  4. [4]

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

    Wenqi Fan, Zihuai Zhao, Jiatong Li, Yunqing Liu, Xiaowei Mei, Yiqi Wang, Jiliang Tang, and Qing Li. Recommender systems in the era of large language models (llms). TKDE, 2024

    work page 2024

  5. [5]

    Chat-rec: Towards interactive and explainable llms-augmented recommender system

    Yunfan Gao, Tao Sheng, Youlin Xiang, Yun Xiong, Haofen Wang, and Jiawei Zhang. Chat-rec: Towards interactive and explainable llms-augmented recommender system. arXiv preprint arXiv:2303.14524, 2023

    arXiv 2023

  6. [6]

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

    Shijie Geng, Shuchang Liu, Zuohui Fu, Yingqiang Ge, and Yongfeng Zhang. Recommendation as language processing (rlp): A unified pretrain, personalized prompt & predict paradigm (p5). In RecSys, pp.\ 299--315, 2022

    work page 2022

  7. [7]

    Neural collaborative filtering

    Xiangnan He, Lizi Liao, Hanwang Zhang, Liqiang Nie, Xia Hu, and Tat-Seng Chua. Neural collaborative filtering. In WWW, pp.\ 173--182, 2017

    work page 2017

  8. [8]

    Lightgcn: Simplifying and powering graph convolution network for recommendation

    Xiangnan He, Kuan Deng, Xiang Wang, Yan Li, Yongdong Zhang, and Meng Wang. Lightgcn: Simplifying and powering graph convolution network for recommendation. In SIGIR, pp.\ 639--648, 2020

    work page 2020

Show all 39 references
  1. [9]

    Lora: Low-rank adaptation of large language models

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. Lora: Low-rank adaptation of large language models. arXiv preprint arXiv:2106.09685, 2021

    2021 arXiv

  2. [10]

    A matrix factorization technique with trust propagation for recommendation in social networks

    Mohsen Jamali and Martin Ester. A matrix factorization technique with trust propagation for recommendation in social networks. In RecSys, pp.\ 135--142, 2010

    2010

  3. [11]

    Adaptive graph contrastive learning for recommendation

    Yangqin Jiang, Chao Huang, and Lianghao Xia. Adaptive graph contrastive learning for recommendation. In KDD, pp.\ 4252--4261, 2023

    2023

  4. [12]

    Diffmm: Multi-modal diffusion model for recommendation

    Yangqin Jiang, Lianghao Xia, Wei Wei, Da Luo, Kangyi Lin, and Chao Huang. Diffmm: Multi-modal diffusion model for recommendation. In ACM MM, pp.\ 7591--7599, 2024

    2024

  5. [13]

    Matrix factorization techniques for recommender systems

    Yehuda Koren, Robert Bell, et al. Matrix factorization techniques for recommender systems. Computer, 0 (8): 0 30--37, 2009

    2009

  6. [14]

    Advances in collaborative filtering

    Yehuda Koren, Steffen Rendle, and Robert Bell. Advances in collaborative filtering. Recommender systems handbook, pp.\ 91--142, 2021

    2021

  7. [15]

    How can recommender systems benefit from large language models: A survey

    Jianghao Lin, Xinyi Dai, Yunjia Xi, Weiwen Liu, Bo Chen, Hao Zhang, Yong Liu, Chuhan Wu, Xiangyang Li, Chenxu Zhu, et al. How can recommender systems benefit from large language models: A survey. TOIS, 2023

    2023

  8. [16]

    Improving graph collaborative filtering with neighborhood-enriched contrastive learning

    Zihan Lin, Changxin Tian, Yupeng Hou, and Wayne Xin Zhao. Improving graph collaborative filtering with neighborhood-enriched contrastive learning. In WWW, pp.\ 2320--2329, 2022

    2022

  9. [17]

    Pre-train, prompt, and recommendation: A comprehensive survey of language modeling paradigm adaptations in recommender systems

    Peng Liu, Lemei Zhang, and Jon Atle Gulla. Pre-train, prompt, and recommendation: A comprehensive survey of language modeling paradigm adaptations in recommender systems. TACL, 11: 0 1553--1571, 2023

    2023

  10. [18]

    Representation learning with large language models for recommendation

    Xubin Ren, Wei Wei, Lianghao Xia, Lixin Su, Suqi Cheng, Junfeng Wang, Dawei Yin, and Chao Huang. Representation learning with large language models for recommendation. In WWW, 2024

    2024

  11. [19]

    Proximal policy optimization algorithms

    John Schulman, Filip Wolski, Prafulla Dhariwal, Alec Radford, and Oleg Klimov. Proximal policy optimization algorithms. arXiv preprint arXiv:1707.06347, 2017

    2017 arXiv

  12. [20]

    Learning to summarize with human feedback

    Nisan Stiennon, Long Ouyang, Jeffrey Wu, Daniel Ziegler, Ryan Lowe, Chelsea Voss, Alec Radford, Dario Amodei, and Paul F Christiano. Learning to summarize with human feedback. NeurIPS, pp.\ 3008--3021, 2020

    2020

  13. [21]

    A survey of collaborative filtering techniques

    Xiaoyuan Su and Taghi M Khoshgoftaar. A survey of collaborative filtering techniques. Advances in artificial intelligence, 2009, 2009

    2009

  14. [22]

    Llama 2: Open foundation and fine-tuned chat models

    Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, et al. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288, 2023

    2023 arXiv

  15. [23]

    Time to shop for valentine's day: Shopping occasions and sequential recommendation in e-commerce

    Jianling Wang, Raphael Louca, Diane Hu, Caitlin Cellier, James Caverlee, and Liangjie Hong. Time to shop for valentine's day: Shopping occasions and sequential recommendation in e-commerce. In WSDM, pp.\ 645--653, 2020

    2020

  16. [24]

    Large language models as data augmenters for cold-start item recommendation

    Jianling Wang, Haokai Lu, James Caverlee, Ed H Chi, and Minmin Chen. Large language models as data augmenters for cold-start item recommendation. In WWW, pp.\ 726--729, 2024

    2024

  17. [25]

    Neural graph collaborative filtering

    Xiang Wang, Xiangnan He, Meng Wang, et al. Neural graph collaborative filtering. In SIGIR, 2019

    2019

  18. [26]

    Llmrec: Large language models with graph augmentation for recommendation

    Wei Wei, Xubin Ren, Jiabin Tang, Qinyong Wang, Lixin Su, Suqi Cheng, Junfeng Wang, Dawei Yin, and Chao Huang. Llmrec: Large language models with graph augmentation for recommendation. In WSDM, pp.\ 806--815, 2024

    2024

  19. [27]

    Mind: A large-scale dataset for news recommendation

    Fangzhao Wu, Ying Qiao, Jiun-Hung Chen, Chuhan Wu, Tao Qi, Jianxun Lian, Danyang Liu, Xing Xie, Jianfeng Gao, Winnie Wu, et al. Mind: A large-scale dataset for news recommendation. In ACL, pp.\ 3597--3606, 2020

    2020

  20. [28]

    Self-supervised graph learning for recommendation

    Jiancan Wu, Xiang Wang, Fuli Feng, Xiangnan He, Liang Chen, Jianxun Lian, and Xing Xie. Self-supervised graph learning for recommendation. In SIGIR, pp.\ 726--735, 2021

    2021

  21. [29]

    Turning clicks into purchases: Revenue optimization for product search in e-commerce

    Liang Wu, Diane Hu, Liangjie Hong, and Huan Liu. Turning clicks into purchases: Revenue optimization for product search in e-commerce. In SIGIR, pp.\ 365--374, 2018

    2018

  22. [30]

    Self-supervised learning for large-scale item recommendations

    Tiansheng Yao, Xinyang Yi, Derek Zhiyuan Cheng, Felix Yu, Ting Chen, Aditya Menon, Lichan Hong, Ed H Chi, Steve Tjoa, Jieqi Kang, et al. Self-supervised learning for large-scale item recommendations. In CIKM, pp.\ 4321--4330, 2021

    2021

  23. [31]

    Are graph augmentations necessary? simple graph contrastive learning for recommendation

    Junliang Yu, Hongzhi Yin, Xin Xia, Tong Chen, Lizhen Cui, and Quoc Viet Hung Nguyen. Are graph augmentations necessary? simple graph contrastive learning for recommendation. In SIGIR, pp.\ 1294--1303, 2022

    2022

  24. [32]

    Self-supervised learning for recommender systems: A survey

    Junliang Yu, Hongzhi Yin, Xin Xia, Tong Chen, Jundong Li, and Zi Huang. Self-supervised learning for recommender systems: A survey. TKDE, 2023

    2023

  25. [33]

    Where to go next for recommender systems? id-vs

    Zheng Yuan, Fajie Yuan, Yu Song, Youhua Li, Junchen Fu, Fei Yang, Yunzhu Pan, and Yongxin Ni. Where to go next for recommender systems? id-vs. modality-based recommender models revisited. In SIGIR, pp.\ 2639--2649, 2023

    2023

  26. [34]

    Deconfounding duration bias in watch-time prediction for video recommendation

    Ruohan Zhan, Changhua Pei, Qiang Su, Jianfeng Wen, Xueliang Wang, Guanyu Mu, Dong Zheng, Peng Jiang, and Kun Gai. Deconfounding duration bias in watch-time prediction for video recommendation. In KDD, pp.\ 4472--4481, 2022

    2022

  27. [35]

    Understanding wechat user preferences and “wow” diffusion

    Fanjin Zhang, Jie Tang, Xueyi Liu, Zhenyu Hou, Yuxiao Dong, Jing Zhang, Xiao Liu, Ruobing Xie, Kai Zhuang, Xu Zhang, et al. Understanding wechat user preferences and “wow” diffusion. TKDE, pp.\ 6033--6046, 2021

    2021

  28. [36]

    Recommendation as instruction following: A large language model empowered recommendation approach

    Junjie Zhang, Ruobing Xie, Yupeng Hou, Xin Zhao, Leyu Lin, and Ji-Rong Wen. Recommendation as instruction following: A large language model empowered recommendation approach. TOIS, 2023

    2023

  29. [37]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  30. [38]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  31. [39]

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.