ELIXIR: Efficient and LIghtweight model for eXplaIning Recommendations

Ben Kabongo; Pirmin Lemberger; Vincent Guigue

REVIEW 4 major objections 5 minor 1 cited by

ELIXIR: Efficient and LIghtweight model for eXplaIning Recommendations

T0 review · 4 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read ELIXIR shows that a frozen 60-million-parameter T5, steered by aspect-aware continuous prompts, generates personalized recommendation reviews that outperform much larger fine-tuned language models on two multi-aspect datasets.

desk verdict Promising architecture, but the headline comparison isn't controlled for the extra aspect supervision ELIXIR receives, and some reported numbers need checking. read the letter →

arxiv 2508.20312 v2 pith:UOBEX4FU submitted 2025-08-27 cs.IR cs.CLcs.LG

Ben Kabongo , Vincent Guigue , Pirmin Lemberger This is my paper · ORCID

classification cs.IRcs.CLcs.LG

keywords recommendersystemsexplanationgenerationreviewaspect-basedrecommendationprompttuningpersonalizedattentionmulti-tasklearninglargelanguagemodels

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

This paper tries to establish that the quality of a generated recommendation explanation depends less on the scale of the language model than on how user and item information is injected into it. Concretely, ELIXIR builds global and aspect-specific user and item representations, weights aspects by a learned personalized attention mechanism, and compresses them into a 50-token continuous prompt for a frozen T5-small. Across TripAdvisor and RateBeer, it reports substantially higher review-generation scores than RNN baselines, an unpretrained transformer, and a fine-tuned GPT-2 model about twice its size. If this holds, it would make lightweight and transparent explainable recommendation practical, since the same prompt machinery also predicts overall and aspect ratings and yields attention weights that can be read as explanations.

What carries the argument

The central object is the personalized continuous prompt $p_{ui} = \psi(u, i, \{a_u, a_i\}_{a\in A})$, a sequence of $\eta = 50$ learnable vectors in the language model's embedding space, generated from global user and item embeddings and aspect-specific MLP projections, weighted by personalized attention coefficients $\alpha^a_u$ and $\alpha^a_i$ computed through query-key dot products. This prompt replaces the usual one- or two-token user and item conditioning used by prior transformers, and because the language model stays frozen, only the prompt generator and rating prediction heads are trained, first on rating losses and then on review negative log-likelihood. The machinery performs three tasks at once: it predicts overall and aspect ratings, it produces the prompt that steers the frozen decoder, and its attention weights can be interpreted as per-aspect importance for each user and item.

What would settle it

Shuffle the per-aspect ratings of a test user-item pair before building the continuous prompt and regenerate the review; if the output stays essentially the same, the aspect-conditioned prompt is not carrying the aspect information, undercutting the paper's central explanation for its gains.

Watch

Extended reading notes

Core claim

ELIXIR's central discovery is that aspect-aware continuous prompting, learned from per-aspect ratings, is what lets a small frozen language model generate personalized reviews. The paper frames the problem as three joint objectives—overall rating prediction, aspect rating prediction, and review likelihood—with shared global and aspect-level embeddings. The generative module receives a continuous prompt computed from those embeddings via nonlinear projections, while the language model (T5-small, 60M parameters) stays frozen during training and only the prompt generator and rating heads are optimized, in a sequential schedule. On TripAdvisor, ELIXIR reaches METEOR 42.75 and BLEU 33.54, against 24.34 and 11.40 for the GPT-2-based PEPLER baseline; on RateBeer the margins are smaller but consistent. The ablation without aspect modeling (ELIXIR-Aspects) drops to 27.64 METEOR on TripAdvisor, close to the best baseline, which the paper reads as evidence that aspect modeling, not model size, is the operative ingredient.

Load-bearing premise

The model needs a reliable aspect-level rating for every user-item interaction, and without those annotations the reported gains in review generation largely disappear.

Editorial extensions

If this is right

A fixed pretrained language model can be steered by a compact learned prompt to produce personalized text, so explainable recommendation does not require fine-tuning large models.
Aspect-supervised representations improve overall rating, aspect rating, and review generation simultaneously, giving richer explanations without extra parameters at inference time.
Prompt length is a genuine hyperparameter: with 50 tokens ELIXIR beats the GPT-2-based PEPLER on TripAdvisor, while with only 2 or 5 tokens it falls far below that baseline.
The learned attention weights align with what users actually mention in their reviews, so the same forward pass that generates text also produces a human-readable justification of the recommendation.
Reliable aspect ratings are the fuel for the whole system; the paper's stated next step is unsupervised aspect extraction so the approach can move beyond annotated datasets.

Reading between the lines

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

The sequential training choice suggests the prompt generator inherits whatever structure the rating module has learned; a testable prediction is that perturbing aspect ratings at inference should shift the aspects mentioned in the generated review before it shifts the overall sentiment.
Because the no-aspect ablation loses most of the advantage, the approach's fate on unlabeled data hinges on the quality of an automatic aspect extractor, and datasets with noisy aspect labels should reproduce the TripAdvisor gap only if the extraction is accurate.
The same personalized-prompt machinery could be lifted to other user-conditioned text tasks such as review summarization, personalized product descriptions, or explanation of search results, whenever structured per-aspect preferences are available.
The sharp jump in quality between 20 and 50 prompt tokens hints at an information bottleneck rather than a smooth improvement, suggesting an information-theoretic study of what the prompt encodes could identify the minimal prompt size for a given task.

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.

Desk Editor's Note

read the letter

Colleague,

The paper is worth a look, but the headline result is not yet a controlled comparison. The authors claim ELIXIR, a frozen T5-small with aspect-aware continuous prompts, beats larger fine-tuned models like PEPLER-GPT2 by huge margins on review generation. The most likely driver is extra supervision: ELIXIR is trained on aspect-level ratings for every interaction, while PEPLER and PETER are not. The ablation ELIXIR-Aspects, which removes aspect modeling, loses most of the gain (METEOR drops from 42.75 to 27.64 on TripAdvisor). It still beats PEPLER on METEOR, so the prompt tuning has some independent value, but the paper's "state-of-the-art" framing overstates the comparison.

What is genuinely good: the architecture is clean, the ablations isolate the contributions of the components, the paper is clearly written, and the code is released. A frozen 60M model steered by a learnable prompt is a practically interesting recipe.

The soft spots are real. First, the paper dismisses aspect-aware Transformer extensions ([5,39,43]) with an unmeasured claim that they "closely mirror" PETER/PEPLER; those are exactly the baselines needed to separate the effect of aspect modeling from the architecture. Second, there are no significance tests or variance estimates, so "significantly outperforms" is not supported. Third, Table 4 lists identical Att2Seq scores for TripAdvisor and RateBeer, which cannot be right and suggests a transcription error. Fourth, Table 2 on RateBeer contradicts the "consistently outperforms all baselines" claim: MF (0.411) and PETER (0.415) beat ELIXIR (0.416) on RMSE. The text admits this but the abstract does not. Fifth, the reliance on aspect annotations is acknowledged in the limitations, but it is a genuine constraint on deployment.

If I were refereeing this, I would ask for the aspect-aware baselines, significance tests, and a corrected Table 2/4 before accepting. The core idea is plausible and the empirical gains are large enough to merit a serious referee.

Recommendation: send to peer review. It is not a desk reject; it needs revision.

Referee Report

4 major / 5 minor

Summary. The paper proposes ELIXIR, a multi-task model for explainable recommendation that jointly predicts overall ratings, aspect-level ratings, and generates personalized review text. The architecture learns global and aspect-specific user/item representations, aggregates them via a personalized attention mechanism, and feeds a continuous prompt into a frozen T5-small language model. Experiments on TripAdvisor and RateBeer compare ELIXIR against matrix factorization, aspect-based recommenders (ALFM, ANR), and review-generation baselines (Att2Seq, NRT, PETER, PEPLER-GPT2), with ablations isolating the contributions of personalized attention, aspect modeling, and global representations. The paper reports large gains in review-generation metrics, especially METEOR and BLEU, and concludes that a 60M-parameter frozen T5 guided by aspect-aware prompts outperforms much larger fine-tuned GPT-2 baselines.

Significance. If the reported gains hold under controlled comparison, the paper would make a useful contribution: it demonstrates that parameter-efficient prompt tuning with aspect-aware continuous prompts can steer a small frozen language model to produce more personalized and higher-quality reviews than full fine-tuning of a larger model. The ablation study is well designed and provides evidence for the individual contributions of aspect modeling and personalized attention. The paper also ships code and reports a concrete hyperparameter analysis (Table 5), which strengthens reproducibility. However, the significance of the central claim depends on resolving the fairness of the baseline comparison: the main baselines do not receive the aspect-level supervision that ELIXIR uses, and the paper dismisses existing aspect-aware Transformer baselines without measurements. The textual claim of consistent and significant superiority is also not fully supported by the tables, since ELIXIR does not beat MF or PETER on RateBeer RMSE.

major comments (4)

[§4.3.3 and §4.5 (Table 4)] The comparison against Transformer baselines is not controlled for the main additional input. ELIXIR is trained with aspect-level rating targets and aspect-aware representations, while PEPLER-GPT2 and PETER are not; Section 4.3.3 excludes the aspect-aware extensions from [5, 39, 43] with the unsupported assertion that their performance 'closely mirrors' PETER/PEPLER. Since the ELIXIR-Aspects ablation still outperforms PEPLER on TripAdvisor (METEOR 27.64 vs 24.34) but the gap widens substantially with aspect supervision (42.75 vs 24.34), the reader cannot tell how much of the headline gain comes from the prompt-tuning architecture and how much from access to fine-grained aspect labels. Please either include one or more aspect-aware Transformer baselines with measurements or explicitly scope the claim to 'among methods without aspect supervision on the generation side.'
[§4.4 and Table 2] The text in §4.4 says ELIXIR 'consistently outperforms all baselines across all metrics,' but Table 2 shows MF achieves RMSE 0.411 on RateBeer versus ELIXIR's 0.416, and PETER achieves 0.415 versus ELIXIR's 0.416. The abstract and conclusion repeat the 'significantly outperforms' wording. This is an internal inconsistency in the manuscript's central performance claim. Please correct the wording, report variance or significance tests, or add a statistical justification for why the differences should be considered meaningful.
[§4.5, Table 4] Att2Seq is reported with identical scores on TripAdvisor and RateBeer for all eight metrics (METEOR 18.611, BLEU 04.690, etc.). Identical values across two datasets that differ in size, domain, and review distribution are implausible and indicate a likely copy-paste error or a failure in the evaluation pipeline. Since the table is the primary evidence for the review-generation claims, please rerun the baseline or correct the numbers and explain why the values coincide.
[§3.3.2 and §4.3.4] The sequential training procedure is described only briefly, and the manuscript does not report how the two training stages interact with validation-based early stopping. In particular, the rating-prediction stage is trained for 50 epochs and the generation stage for 50 epochs, but there is no analysis of how sensitive the final review-generation quality is to the number of epochs in each stage or to the choice of alpha in Eq. (11). Given that the ablation results hinge on the quality of the learned aspect representations from the first stage, a sensitivity analysis for alpha and for the split point of the sequential training would strengthen the reproducibility of the reported results.

minor comments (5)

[§4.3.1] The table header 'Statistics after filtering are shown in brackets' and the entries like '40 213 (8 384)' are difficult to parse; please present pre- and post-filtering counts in separate columns or more clearly label the bracketed numbers.
[§4.3.1] There is a typo in 'real-word multi-aspect datasets'; it should be 'real-world.'
[§4.3.3] When describing PEPLER, the paper says it 'employs the pre-trained GPT-2 in a fine-tuning approach,' but in the setup it states PEPLER uses GPT-2 (124M); please clarify whether the reported baseline uses GPT-2-small and confirm that the evaluation protocol (including prompt construction and decoding) matches the original PEPLER paper.
[Table 3] The table reports standard deviations across aspects, but the accompanying text says ELIXIR 'consistently and significantly outperforms' without any statistical test. Please either report significance tests across runs or soften the 'significantly' wording.
[§4.6 and Figure 3] The qualitative claim that aspect representations 'allow a coherent separation of users according to the aspects' is based on a single t-SNE-style projection; please provide quantitative clustering metrics or a second domain example to support the interpretability claim.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the claimed gains rest on empirical comparisons against external baselines, and no modeling step reduces to its own inputs by construction.

full rationale

ELIXIR's central claims are empirical: the model is trained and evaluated against external baselines (Att2Seq, NRT, PETER, PEPLER) on public TripAdvisor and RateBeer datasets. The modeling equations (2)-(8) define user/item global and aspect representations, attention weights, rating predictors, and a prompt-conditioned review likelihood; none of these definitions presuppose the target result. The review-generation loss (12) is standard negative log-likelihood on held-out reviews, and the reported METEOR/BLEU/ROUGE/BERTScore numbers are computed on test-set outputs, so the headline gains are not fitted quantities renamed as predictions. The ablations (ELIXIR-Attention, ELIXIR-Aspects, ELIXIR-Global) compare design variants independently from the central claim. The attention-importance analysis in Section 4.7 and Table 6 is an illustrative post-hoc correlation, not a derivation. There are no self-citations: the reference list contains no work by the present authors, and no uniqueness theorem is imported. The only flagged weakness is experimental fairness rather than circularity: Section 4.3.3 excludes the aspect-aware Transformer extensions [5,39,43] with the unsupported assertion that their performance 'closely mirrors' PETER/PEPLER, and Section 5 concedes reliance on aspect annotations, which restricts applicability; these are validity and robustness concerns. The derivation chain itself is self-contained, so the circularity score is 0.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The model is a supervised learning system. Its performance claims rest on tuned hyperparameters (eta, alpha, dimension) and domain assumptions: aspect labels must exist, automatic metrics must measure explanation quality, the evaluation split must reflect deployment, and a small frozen LM must be sufficient. No new physical or conceptual entities are introduced.

free parameters (3)

eta (number of prompt tokens) = 50
Set after preliminary experiments on the validation set (Section 4.3.4). Review generation quality scales with eta (Table 5), so the SOTA claim depends on this choice.
alpha (rating loss weight) = 1/(|A|+1)
Chosen as the best balance between overall and aspect rating losses (Section 4.3.4); affects rating prediction and the representations used for prompts.
model dimension d = 256
Dimension of global and aspect representations, fixed in Section 4.3.4 without sensitivity analysis.

assumptions (4)

domain assumption Aspect-level ratings are available for every user-item interaction or can be extracted reliably with ABSA.
The architecture's aspect modeling, aspect rating loss, and personalized attention all depend on aspect labels (Section 3.1). The paper acknowledges this in Section 5.
domain assumption Automatic text metrics (BLEU, ROUGE, METEOR, BERTScore) are valid proxies for explanation quality.
These are the only evaluation measures for review generation (Section 4.3.2); no human evaluation or factuality check is performed.
domain assumption The random 80:10:10 split is i.i.d., so user/item representations learned from training interactions generalize to the test set.
Standard assumption in recommender evaluations; no temporal or interaction-order information is used (Section 4.3.4).
domain assumption A frozen T5-small contains enough linguistic knowledge to generate fluent, personalized reviews when given a 50-token continuous prompt.
The method relies on prompt tuning to access pre-trained knowledge (Sections 3.2.4 and 4.5).

how reviews work

0 comments

Cite this review

Pith. "Pith review of ELIXIR: Efficient and LIghtweight model for eXplaIning Recommendations." pith.science (2026). https://pith.science/paper/UOBEX4FU

@misc{pith2026250820312,
  author       = {Pith},
  title        = {Pith review of: ELIXIR: Efficient and LIghtweight model for eXplaIning Recommendations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UOBEX4FU}},
  note         = {Machine review of arXiv:2508.20312}
}

read the original abstract

Collaborative filtering drives many successful recommender systems but struggles with fine-grained user-item interactions and explainability. As users increasingly seek transparent recommendations, generating textual explanations through language models has become a critical research area. Existing methods employ either RNNs or Transformers. However, RNN-based approaches fail to leverage the capabilities of pre-trained Transformer models, whereas Transformer-based methods often suffer from suboptimal adaptation and neglect aspect modeling, which is crucial for personalized explanations. We propose ELIXIR (Efficient and LIghtweight model for eXplaIning Recommendations), a multi-task model combining rating prediction with personalized review generation. ELIXIR jointly learns global and aspect-specific representations of users and items, optimizing overall rating, aspect-level ratings, and review generation, with personalized attention to emphasize aspect importance. Based on a T5-small (60M) model, we demonstrate the effectiveness of our aspect-based architecture in guiding text generation in a personalized context, where state-of-the-art approaches exploit much larger models but fail to match user preferences as well. Experimental results on TripAdvisor and RateBeer demonstrate that ELIXIR significantly outperforms strong baseline models, especially in review generation.

Figures

Figures reproduced from arXiv: 2508.20312 by the authors.

**Figure 2.** ELIXIR consists of two modules: (a) the rating prediction module on the left and (b) the personalized review generation [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗

**Figure 3.** Projection and clustering of user aspect representa [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

Rank, Don't Generate: Statement-level Ranking for Explainable Recommendation
cs.IR 2026-04 unverdicted novelty 7.0 of 10

The work reframes explainable recommendation as statement-level ranking, introduces the StaR benchmark from Amazon reviews, and finds popularity baselines outperforming SOTA models in item-level personalized ranking.

Reference graph

Works this paper leans on

59 extracted references · 27 canonical work pages · cited by 1 Pith paper

[1]

Satanjeev Banerjee and Alon Lavie. 2005. METEOR: An automatic metric for MT evaluation with improved correlation with human judgments. InProceedings of the acl workshop on intrinsic and extrinsic evaluation measures for machine translation and/or summarization. 65–72

2005
[2]

Konstantin Bauman, Bing Liu, and Alexander Tuzhilin. 2017. Aspect based recommendations: Recommending items with the most valuable aspects based on user reviews. InProceedings of the 23rd ACM SIGKDD international conference on knowledge discovery and data mining. 717–725

work page 2017
[3]

Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. 2020. Language models are few-shot learners.Advances in neural information processing systems33 (2020), 1877–1901

2020
[4]

Xu Chen, Zheng Qin, Yongfeng Zhang, and Tao Xu. 2016. Learning to rank features for recommendation over multiple categories. InProceedings of the 39th International ACM SIGIR conference on Research and Development in Information Retrieval. 305–314

work page 2016
[5]

Hao Cheng, Shuo Wang, Wensheng Lu, Wei Zhang, Mingyang Zhou, Kezhong Lu, and Hao Liao. 2023. Explainable recommendation with personalized review retrieval and aspect learning.arXiv preprint arXiv:2306.12657(2023)

work page Pith review arXiv 2023
[6]

Zhiyong Cheng, Ying Ding, Xiangnan He, Lei Zhu, Xuemeng Song, and Mohan S Kankanhalli. 2018. Aˆ 3NCF: An Adaptive Aspect Attention Model for Rating Prediction.. InIJCAI. 3748–3754

work page 2018
[7]

Zhiyong Cheng, Ying Ding, Lei Zhu, and Mohan Kankanhalli. 2018. Aspect-aware latent factor model: Rating prediction with ratings and reviews. InProceedings of the 2018 world wide web conference. 639–648

work page 2018
[8]

Jin Yao Chin, Kaiqi Zhao, Shafiq Joty, and Gao Cong. 2018. ANR: Aspect-based neural recommender. InProceedings of the 27th ACM International conference on information and knowledge management. 147–156

work page 2018

Show all 59 references

[9]

Felipe Costa, Sixun Ouyang, Peter Dolog, and Aonghus Lawlor. 2018. Automatic generation of natural language explanations. InCompanion Proceedings of the 23rd International Conference on Intelligent User Interfaces. 1–2

2018
[10]

Li Dong, Shaohan Huang, Furu Wei, Mirella Lapata, Ming Zhou, and Ke Xu
[11]

Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Amy Yang, Angela Fan, et al. 2024. The llama 3 herd of models.arXiv preprint arXiv:2407.21783(2024)

2024 arXiv
[12]

Gayatree Ganu, Noemie Elhadad, and Amélie Marian. 2009. Beyond the stars: Improving rating predictions using review text content.. InWebDB, Vol. 9. 1–6

2009
[13]

Yingqiang Ge, Shuchang Liu, Zuohui Fu, Juntao Tan, Zelong Li, Shuyuan Xu, Yunqi Li, Yikun Xian, and Yongfeng Zhang. 2024. A survey on trustworthy recommender systems.ACM Transactions on Recommender Systems3, 2 (2024), 1–68

2024
[14]

Shijiang Geng, Shuai Liu, Zhiyong Fu, Yong Ge, and Yu Zhang. 2023. Rec- ommendation as Language Processing (RLP): A Unified Pretrain, Personalized Prompt & Predict Paradigm (P5).arXiv preprint arXiv:2203.13366(Jan 2023). doi:10.48550/arXiv.2203.13366

2023 doi arXiv
[15]

Xinyu Guan, Zhiyong Cheng, Xiangnan He, Yongfeng Zhang, Zhibo Zhu, Qinke Peng, and Tat-Seng Chua. 2019. Attentive aspect modeling for review-aware recommendation.ACM Transactions on Information Systems (TOIS)37, 3 (2019), 1–27

2019
[16]

Zeyu Han, Chao Gao, Jinyang Liu, Jeff Zhang, and Sai Qian Zhang. 2024. Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey. Transactions on Machine Learning Research(2024)

2024
[17]

Emrul Hasan, Mizanur Rahman, Chen Ding, Jimmy Xiangji Huang, and Shaina Raza. 2024. Review-based Recommender Systems: A Survey of Approaches, Challenges and Future Perspectives.arXiv preprint arXiv:2405.05562(2024)

2024 arXiv
[18]

Xiangnan He, Tao Chen, Min-Yen Kan, and Xiao Chen. 2015. Trirank: Review- aware explainable recommendation by modeling aspects. InProceedings of the 24th ACM international on conference on information and knowledge management. 1661–1670

2015
[19]

Xiangnan He, Lizi Liao, Hanwang Zhang, Liqiang Nie, Xia Hu, and Tat-Seng Chua. 2017. Neural collaborative filtering. InProceedings of the 26th international conference on world wide web. 173–182

2017
[20]

Umair Javed, Kamran Shaukat, Ibrahim A Hameed, Farhat Iqbal, Talha Mahboob Alam, and Suhuai Luo. 2021. A review of content-based and context-based recommendation systems.International Journal of Emerging Technologies in Learning (iJET)16, 3 (2021), 274–306

2021
[21]

Diederik P Kingma. 2014. Adam: A method for stochastic optimization.arXiv preprint arXiv:1412.6980(2014)

2014 arXiv
[22]

Yehuda Koren. 2009. Collaborative filtering with temporal dynamics. InProceed- ings of the 15th ACM SIGKDD international conference on Knowledge discovery and data mining. 447–456

2009
[23]

Yehuda Koren, Robert Bell, and Chris Volinsky. 2009. Matrix factorization tech- niques for recommender systems.Computer42, 8 (2009), 30–37

2009
[24]

Yehuda Koren, Steffen Rendle, and Robert Bell. 2021. Advances in collaborative filtering.Recommender systems handbook(2021), 91–142

2021
[25]

Brian Lester, Rami Al-Rfou, and Noah Constant. 2021. The power of scale for parameter-efficient prompt tuning.arXiv preprint arXiv:2104.08691(2021)

2021 arXiv
[26]

Lei Li, Yongfeng Zhang, and Li Chen. 2021. Personalized transformer for explain- able recommendation.arXiv preprint arXiv:2105.11601(2021)

2021 arXiv
[27]

Lei Li, Yongfeng Zhang, and Li Chen. 2023. Personalized prompt learning for explainable recommendation.ACM Transactions on Information Systems41, 4 (2023), 1–26

2023
[28]

Piji Li, Zihao Wang, Zhaochun Ren, Lidong Bing, and Wai Lam. 2017. Neural rating regression with abstractive tips generation for recommendation. InProceed- ings of the 40th International ACM SIGIR conference on Research and Development in Information Retrieval. 345–354

2017
[29]

Chin-Yew Lin. 2004. Rouge: A package for automatic evaluation of summaries. InText summarization branches out. 74–81

2004
[30]

2022.Sentiment analysis and opinion mining

Bing Liu. 2022.Sentiment analysis and opinion mining. Springer Nature

2022
[31]

Pasquale Lops, Marco De Gemmis, and Giovanni Semeraro. 2011. Content- based recommender systems: State of the art and trends.Recommender systems handbook(2011), 73–105

2011
[32]

Julian McAuley, Jure Leskovec, and Dan Jurafsky. 2012. Learning attitudes and attributes from multi-aspect reviews. In2012 IEEE 12th International Conference on Data Mining. IEEE, 1020–1025

2012
[33]

Andriy Mnih and Russ R Salakhutdinov. 2007. Probabilistic matrix factorization. Advances in neural information processing systems20 (2007)

2007
[34]

Vinod Nair and Geoffrey E Hinton. 2010. Rectified linear units improve re- stricted boltzmann machines. InProceedings of the 27th international conference on machine learning (ICML-10). 807–814

2010
[35]

Jianmo Ni, Jiacheng Li, and Julian McAuley. 2019. Justifying recommendations using distantly-labeled reviews and fine-grained aspects. InProceedings of the 2019 conference on empirical methods in natural language processing and the 9th international joint conference on natural...

2019
[36]

Jianmo Ni and Julian McAuley. 2018. Personalized review generation by expand- ing phrases and attending on aspect-aware representations. InProceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers). 706–711

2018
[37]

Harris Papadakis, Antonis Papagrigoriou, Costas Panagiotakis, Eleftherios Kos- mas, and Paraskevi Fragopoulou. 2022. Collaborative filtering recommender systems taxonomy.Knowledge and Information Systems64, 1 (2022), 35–74

2022
[38]

Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002. Bleu: a method for automatic evaluation of machine translation. InProceedings of the 40th annual meeting of the Association for Computational Linguistics. 311–318

2002
[39]

Jakub Raczyński, Mateusz Lango, and Jerzy Stefanowski. 2023. The Problem of Coherence in Natural Language Explanations of Recommendations. InECAI

2023
[40]

Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. 2019. Language models are unsupervised multitask learners.OpenAI blog1, 8 (2019), 9. https://cdn.openai.com/better-language-models/language_models_ are_unsupervised_multitask_learners.pdf

2019
[41]

Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J Liu. 2020. Exploring the limits of transfer learning with a unified text-to-text transformer.Journal of machine learning research21, 140 (2020), 1–67

2020
[42]

Steffen Rendle, Walid Krichene, Li Zhang, and John Anderson. 2020. Neural collaborative filtering vs. matrix factorization revisited. InProceedings of the 14th ACM Conference on Recommender Systems. 240–248

2020
[43]

Ryotaro Shimizu, Takashi Wada, Yu Wang, Johannes Kruse, Sean O’Brien, Sai HtaungKham, Linxin Song, Yuya Yoshikawa, Yuki Saito, Fugee Tsung, et al
[44]

Jie Shuai, Kun Zhang, Le Wu, Peijie Sun, Richang Hong, Meng Wang, and Yong Li
[45]

Peijie Sun, Le Wu, Kun Zhang, Yu Su, and Meng Wang. 2021. An unsupervised aspect-aware recommendation model with explanation text generation.ACM Transactions on Information Systems (TOIS)40, 3 (2021), 1–29

2021
[46]

Gomez, Lukasz Kaiser, and Illia Polosukhin

Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. 2017. Attention Is All You Need. arXiv:1706.03762 [cs.CL] https://arxiv.org/abs/1706.03762

2017 arXiv
[47]

Hongning Wang, Yue Lu, and Chengxiang Zhai. 2010. Latent aspect rating analysis on review text data: a rating regression approach. InProceedings of the 16th ACM SIGKDD international conference on Knowledge discovery and data mining. 783–792

2010
[48]

Zhouhang Xie, Sameer Singh, Julian McAuley, and Bodhisattwa Prasad Majumder
[49]

Tianyi Zhang, Varsha Kishore, Felix Wu, Kilian Q Weinberger, and Yoav Artzi. 2019. Bertscore: Evaluating text generation with bert.arXiv preprint arXiv:1904.09675(2019)

2019 arXiv
[50]

Wenxuan Zhang, Yang Deng, Xin Li, Yifei Yuan, Lidong Bing, and Wai Lam
[51]

Wenxuan Zhang, Xin Li, Yang Deng, Lidong Bing, and Wai Lam. 2022. A sur- vey on aspect-based sentiment analysis: Tasks, methods, and challenges.IEEE Transactions on Knowledge and Data Engineering35, 11 (2022), 11019–11038

2022
[52]

Yongfeng Zhang, Xu Chen, et al. 2020. Explainable recommendation: A survey and new perspectives.Foundations and Trends®in Information Retrieval14, 1 (2020), 1–101

2020
[53]

InProceedings of the AAAI Conference on Artificial Intelligence, Vol

Factual and informative review generation for explainable recommendation. InProceedings of the AAAI Conference on Artificial Intelligence, Vol. 37. 13816– 13824
[59]

Yongfeng Zhang, Guokun Lai, Min Zhang, Yi Zhang, Yiqun Liu, and Shaoping Ma. 2014. Explicit factor models for explainable recommendation based on phrase-level sentiment analysis. InProceedings of the 37th international ACM SIGIR conference on Research & development in informat...

2014
[2017]

InProceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers

Learning to generate product reviews from attributes. InProceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers. 623–632
[2021]

Aspect sentiment quad prediction as paraphrase generation.arXiv preprint arXiv:2110.00796(2021)

2021 arXiv
[2022]

InProceedings of the 45th international ACM SIGIR conference on research and development in information retrieval

A review-aware graph contrastive learning framework for recommendation. InProceedings of the 45th international ACM SIGIR conference on research and development in information retrieval. 1283–1293
[2023]

IOS Press, 1922–1929

1922
[2024]

Disentangling Likes and Dislikes in Personalized Generative Explainable Recommendation.arXiv preprint arXiv:2410.13248(2024)

2024 arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.

[1] [1]

Satanjeev Banerjee and Alon Lavie. 2005. METEOR: An automatic metric for MT evaluation with improved correlation with human judgments. InProceedings of the acl workshop on intrinsic and extrinsic evaluation measures for machine translation and/or summarization. 65–72

2005

[2] [2]

Konstantin Bauman, Bing Liu, and Alexander Tuzhilin. 2017. Aspect based recommendations: Recommending items with the most valuable aspects based on user reviews. InProceedings of the 23rd ACM SIGKDD international conference on knowledge discovery and data mining. 717–725

work page 2017

[3] [3]

Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. 2020. Language models are few-shot learners.Advances in neural information processing systems33 (2020), 1877–1901

2020

[4] [4]

Xu Chen, Zheng Qin, Yongfeng Zhang, and Tao Xu. 2016. Learning to rank features for recommendation over multiple categories. InProceedings of the 39th International ACM SIGIR conference on Research and Development in Information Retrieval. 305–314

work page 2016

[5] [5]

Hao Cheng, Shuo Wang, Wensheng Lu, Wei Zhang, Mingyang Zhou, Kezhong Lu, and Hao Liao. 2023. Explainable recommendation with personalized review retrieval and aspect learning.arXiv preprint arXiv:2306.12657(2023)

work page Pith review arXiv 2023

[6] [6]

Zhiyong Cheng, Ying Ding, Xiangnan He, Lei Zhu, Xuemeng Song, and Mohan S Kankanhalli. 2018. Aˆ 3NCF: An Adaptive Aspect Attention Model for Rating Prediction.. InIJCAI. 3748–3754

work page 2018

[7] [7]

Zhiyong Cheng, Ying Ding, Lei Zhu, and Mohan Kankanhalli. 2018. Aspect-aware latent factor model: Rating prediction with ratings and reviews. InProceedings of the 2018 world wide web conference. 639–648

work page 2018

[8] [8]

Jin Yao Chin, Kaiqi Zhao, Shafiq Joty, and Gao Cong. 2018. ANR: Aspect-based neural recommender. InProceedings of the 27th ACM International conference on information and knowledge management. 147–156

work page 2018