Pith. sign in

REVIEW 5 major objections 5 minor 96 references

From Reviews to Dialogues: Active Synthesis for Zero-Shot LLM-based Conversational Recommender System

T0 review · 5 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read The paper claims that a conversational recommender can be trained without in-domain dialogue data by actively selecting reviews and metadata and converting them into synthetic dialogues with a teacher LLM.

desk verdict The core idea is useful and the full-SFT results show promise, but the paper's 'consistent improvement' claim is contradicted by its own Table 1 and the experimental details are too incomplete to reproduce. read the letter →

arxiv 2504.15476 v1 pith:3QNKTWBJ submitted 2025-04-21 cs.IR

Rohan Surana , Junda Wu , Zhouhang Xie , Yu Xia , Harald Steck , Dawen Liang , Nathan Kallus , Julian McAuley This is my paper · ORCID
classification cs.IR
keywords conversationalrecommendersystemsactivelearningsyntheticdataaugmentationlargelanguagemodelszero-shotrecommendationscarcitysupervisedfine-tuning
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

The paper tries to establish that a conversational recommender can be built for a domain with no conversational data at all, by converting ordinary item reviews, metadata, and user-item interaction signals into synthetic training dialogues. The conversion is not blind: an active-learning step selects the most informative seed samples before a black-box teacher LLM writes queries and answers, and the resulting dialogues are used to fine-tune a small, internally manageable language model. This matters because many practical deployments cannot call large external models at serving time and lack domain-specific dialogue logs, so a reliable way to synthesize those logs from cheap non-conversational data would remove a central bottleneck. On two standard benchmarks, the paper reports that models trained on its actively synthesized data consistently outperform both zero-shot prompting and fine-tuning on naively generated LLM data, and that the synthetic data can complement or even beat small real dialogue training sets when domain coverage is thin.

What carries the argument

The load-bearing object is the active data augmentation loop, which alternates sample selection with synthetic dialogue generation. Selection uses two information measures over review embeddings: Jensen-Shannon information, built from K-means cluster softmax distributions, entropy, and Jensen-Shannon divergence from the already-selected set, to reward diversity and uncertainty; and Fisher information, computed as $\log(1+\mathbf{x}_i^\top \Lambda_t \mathbf{x}_i)$ with a Sherman-Morrison update of the inverse covariance, to reward samples that most influence the fine-tuning objective. Selected reviews are fed to a teacher LLM along with style templates to produce synthetic user queries, and the same teacher LLM produces the 20-item ground-truth list for each query; the resulting pairs are used for supervised fine-tuning of a smaller language model. The active selection is what distinguishes the approach from naive prompting, because it concentrates the teacher LLM's query budget on diverse and informative seeds.

What would settle it

Run the full pipeline in a new domain whose items appeared after the teacher LLM's training cutoff, where the teacher's recommendations are known to be unreliable, and compare the fine-tuned student against zero-shot prompting on queries with human-verified ground-truth items; if the student does not beat zero-shot, or if its ranking aligns with teacher errors rather than human judgments, the central claim fails.

Watch

Extended reading notes

Core claim

The paper's central claim is that in-domain conversation logs are not a prerequisite for a strong conversational recommender. Starting from a seed set of reviews, item metadata, and collaborative-filtering features, the framework scores every seed with information-theoretic measures, picks the highest-scoring samples under a budget, and prompts a teacher LLM to turn each selected review into a natural user query together with a list of twenty recommended items that serve as ground truth. A smaller language model is then supervised fine-tuned on these synthetic query–recommendation pairs. The experiments report consistent Recall and NDCG gains over zero-shot prompting and over fine-tuning on LLM-generated data without active selection, across several small backbone models, and show that metadata and collaborative signals further improve selection. When real in-domain dialogues are scarce, the synthetic data can outperform training on the small real set and improve it when combined, while supplementing an already well-covered set can add noise.

Load-bearing premise

The paper treats the teacher LLM's own recommended-item lists as correct labels for the synthetic queries, so if those recommendations are systematically biased or unrepresentative of real user preferences, the fine-tuned student learns a faulty recommender and the reported gains may not transfer to genuinely new domains.

Editorial extensions

If this is right

  • A domain with no conversation logs can still get a working conversational recommender by feeding reviews, metadata, and interaction signals through the active-synthesis pipeline.
  • Active seed selection makes synthetic data generation cost-effective: models reach higher Recall and NDCG with fewer teacher-LLM calls than random seed selection.
  • Rich seed signals translate into better synthetic dialogues; adding metadata or collaborative-filtering features to selection improves downstream recommendation.
  • In low-coverage domains, synthetic dialogues can substitute for or complement small real dialogue sets, whereas adding synthetic data to an already well-covered set can slightly hurt performance.

Reading between the lines

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

  • Because active selection favors diverse, uncertain, or high-influence samples, any systematic bias in the teacher LLM's recommendations may be amplified rather than diluted; an extension would add human verification of synthetic labels into the selection loop.
  • The same mechanism should apply to non-movie domains such as books, fashion, or health products, but only where the teacher LLM can rank items plausibly; a cross-domain audit would test whether the gains hold as teacher competence drops.
  • The query-budget savings suggest a practical deployment recipe: keep a small internal model, buy a limited number of teacher-LLM calls per domain, and regenerate the synthetic set periodically as the item catalog changes.
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

5 major / 5 minor

Summary. The paper proposes an active data augmentation framework for zero-shot conversational recommender systems (CRS). Starting from publicly available non-conversational data (item metadata, reviews, and collaborative signals), the method selects supposedly informative seed samples via Jensen-Shannon and Fisher information criteria, prompts a black-box LLM (GPT-4o) to convert those samples into synthetic conversational queries, and then uses the same LLM to generate the ground-truth movie lists for those queries. The resulting synthetic dataset is used to fine-tune smaller language models (Llama3.2-1B, Llama3.2-3B, Gemma2-2B) with either LoRA or full SFT. The paper evaluates on ReDial and INSPIRED, reporting Recall@k and NDCG@k, and claims that models trained on the actively synthesized data consistently outperform both zero-shot and GPT-generated baselines, that active selection reduces API cost, and that metadata and collaborative signals improve selection. It also compares synthetic data with in-domain conversational data (RQ5). The central claim is that the framework provides a general route to building CRS in no-resource or low-resource domains by converting non-conversational data into effective training dialogues.

Significance. If the central claim holds, the framework is practically valuable: it addresses a real bottleneck in CRS research and deployment, namely the scarcity and cost of domain-specific conversational data, while allowing organizations to use smaller, internally managed models rather than black-box APIs. The evaluation on external benchmarks (ReDial and INSPIRED) is a genuine strength and avoids the most direct form of circularity. The idea of combining active selection with synthetic data generation is also well motivated and has clear potential for cost reduction. However, the significance is conditional on the evidence: the headline claim of consistent improvement is not actually supported by Table 1, several important comparisons lack statistical backing, and the synthetic labels are produced by the same teacher model that authors the queries, which is a real correctness risk for generalization to genuinely new domains. The framework is plausible and the reported results are promising in several configurations, but the paper currently overstates what the evidence demonstrates.

major comments (5)
  1. [Section 4.3, Table 1] The claim that "Models trained using our actively synthesized conversational data consistently outperform both zero-shot and GPT-generated baselines" is directly contradicted by Table 1. For example, Llama3-1B LoRA-SFT on INSPIRED achieves R@1=1.9±0.94 vs. GPT-Generated 4.27±1.39, and R@5=6.64 vs. 9.48; Llama3-3B LoRA-SFT on INSPIRED R@20=12.80 vs. 13.27; Gemma2-2B LoRA-SFT on ReDial R@20=15.73 vs. 17.27. Several rows show only small or negative differences. To support the central claim, the paper needs paired significance tests across multiple seeds, error bars on the headline comparisons, and a careful restatement of the claim that distinguishes where active synthesis helps, where it is neutral, and where it hurts.
  2. [Sections 4.4-4.6, Figures 3-7] The claims that JS and Fisher selection "improve recall more efficiently than random sampling" and that metadata/collaborative-signal variants "consistently outperform" semantic-only and random baselines are not quantitatively supported: Figures 3-7 show single curves without error bars, confidence intervals, or significance tests. The text reports no number of random seeds, no variance, and no test statistics for these comparisons. Since random sampling is acknowledged to be a strong baseline, the paper should report seeded repetitions and statistical tests before asserting that the observed differences are reliable.
  3. [Section 3.3, Movie Recommendations Prompt] The synthetic ground-truth movie lists are generated by the same teacher LLM (GPT-4o) that writes the synthetic queries. This means the student is trained on teacher-provided labels, not on verified human preferences. Evaluation on ReDial and INSPIRED mitigates circularity for those two domains, but it does not validate the framework for the claimed transfer to "no-resource" or "rapidly evolving" domains. The paper should either validate the synthetic labels against real user-item interactions or human annotations in a held-out domain, or explicitly discuss this limitation and its implications for the generalization claim.
  4. [Section 3.2.1, Eq. (1)] The entropy term is defined as H(p_i) = sum_j p_ij log p_ij, which is the negative of the standard Shannon entropy. Maximizing this quantity selects samples whose cluster-membership distribution is most concentrated (low entropy), contrary to the stated intention of choosing "the review with maximum entropy." This sign error changes the behavior of the JS-based active selection strategy. The formula should be corrected to H(p_i) = -sum_j p_ij log p_ij, or the selection rule should be changed accordingly.
  5. [Section 4.2.2 and Algorithm 1] Key experimental details necessary for reproducing the main comparison are missing: the exact seed dataset and preprocessing used, the source and number of query templates, the total number of synthetic samples generated, the values of hyperparameters lambda and K, the number of active-selection iterations, and the number of random seeds. Algorithm 1 also uses B ambiguously: it appears both as a selection budget in Section 3.2 and as the number of active-selection iterations in the loop. Without these details, the reported improvements cannot be independently reproduced or compared across configurations.
minor comments (5)
  1. [Section 1] The phrase "back-box LLMs" should read "black-box LLMs".
  2. [Table 1 and text] The model name is spelled inconsistently as both "LlaMA3" and "Llama3" across the table and text; the spelling should be unified.
  3. [Algorithm 2, line 4] The notation "{r1, r2, r3} ⊆ ∇" uses an undefined symbol ∇; this should be a subset of the available reviews.
  4. [Figure 8] The RQ5 comparison is presented only as bar charts without numeric values or error bars, so the claims that Synth outperforms INS on INSPIRED and that R+O degrades ReDial performance cannot be assessed quantitatively.
  5. [Section 4.2.2] The line "batch size is set to 32, with gradient accumulation steps of 8, resulting in an effective minibatch size of 4" is confusing as written; presumably the effective batch size after accumulation should be 256, and the wording should be clarified.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the core improvement claim is tested on external human-annotated benchmarks; the self-generated synthetic labels are a validity limitation, not a derivation-level tautology.

full rationale

The paper's central claim (Section 4.3) is that models fine-tuned on actively synthesized conversational data outperform zero-shot and GPT-generated baselines. This is an empirical comparison conducted on the external, human-annotated ReDial and INSPIRED benchmarks. The metric definitions and evaluation protocol follow [20], a prior work by overlapping authors, but the benchmark labels are not produced by the method. The synthetic training tuples (s_p,k, M_p,k) are generated by the same teacher LLM (Section 3.3, Query Generation Prompt and Movie Recommendations Prompt), which is a validity limitation: the training signal is a distillation of GPT-4o rather than verified user ground truth. However, this is not a derivation-level circularity, because the paper never claims to derive ReDial/INSPIRED performance from the teacher's outputs by construction. No parameter is fitted to the evaluation set and then reported as a prediction. The active-selection strategies (JS/Fisher) are compared on the benchmarks and all variants are reported, rather than selecting the best and calling it a predicted result. The self-citations (especially [20] for the LLM-generated ground-truth protocol) are methodological conventions, not load-bearing uniqueness arguments. The paper's contribution is therefore self-contained in the sense required by the circularity analysis, even though the synthetic-label validity remains an independent methodological concern.

Assumptions & free parameters 5 free parameters · 6 assumptions · 0 invented entities

The paper introduces no new physical or conceptual entities. It relies on a set of domain assumptions about the quality of teacher-generated labels, the representativeness of query templates, and the transferability of synthetic training data. The main free parameters are the active-selection hyperparameters, which are not reported.

free parameters (5)
  • lambda (JS-Fisher balance) = not reported
    Section 3.2.1 defines the selection score as H(p_i) + lambda * JS(...); lambda is a hand-tuned balance coefficient with no reported value or tuning procedure.
  • K (number of K-means clusters) = not reported
    Section 3.2.1 partitions the embedding space into K clusters; K is a free choice and is not specified.
  • LLM generation temperature = 0.8
    Section 4.2.2 sets the GPT-4o generation temperature to 0.8; this is a hand-picked hyperparameter that influences synthetic data diversity.
  • query templates sampled per item = 5
    Section 3.3 randomly selects five query templates per item; this is an arbitrary choice affecting style diversity.
  • reviews sampled per item = 3
    Section 3.3 samples three reviews per generated query; this is an arbitrary choice affecting content coverage.
assumptions (6)
  • domain assumption Teacher LLM outputs are valid proxy labels
    Section 3.3 generates ground-truth movie lists from the same LLM that writes the queries; no validation against human preferences is provided.
  • domain assumption Query templates are representative without leaking test data
    Section 3.3 uses Qtmp to style synthetic queries, but the paper never states where Qtmp comes from.
  • domain assumption Seed data covers the evaluation domain
    The pipeline uses Amazon Reviews '23 plus collaborative signals and tests on ReDial and INSPIRED; domain overlap between the seed data and benchmarks is assumed.
  • domain assumption Language model embeddings capture review informativeness
    Section 3.2 selects samples based on the last-layer hidden states of the target LM, assuming these embeddings reflect useful semantic and collaborative properties.
  • domain assumption SFT on synthetic dialogues transfers to real conversations
    Section 3.4 assumes that training on synthetic (query, movie list) pairs improves real conversational recommendation without catastrophic distribution shift.
  • standard math Standard linear algebra and probability results
    Sherman-Morrison update and Jensen-Shannon divergence are standard results used without proof.

how reviews work

0 comments
Cite this review

Pith. "Pith review of From Reviews to Dialogues: Active Synthesis for Zero-Shot LLM-based Conversational Recommender System." pith.science (2026). https://pith.science/paper/3QNKTWBJ

@misc{pith2026250415476,
  author       = {Pith},
  title        = {Pith review of: From Reviews to Dialogues: Active Synthesis for Zero-Shot LLM-based Conversational Recommender System},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3QNKTWBJ}},
  note         = {Machine review of arXiv:2504.15476}
}
read the original abstract

Conversational recommender systems (CRS) typically require extensive domain-specific conversational datasets, yet high costs, privacy concerns, and data-collection challenges severely limit their availability. Although Large Language Models (LLMs) demonstrate strong zero-shot recommendation capabilities, practical applications often favor smaller, internally managed recommender models due to scalability, interpretability, and data privacy constraints, especially in sensitive or rapidly evolving domains. However, training these smaller models effectively still demands substantial domain-specific conversational data, which remains challenging to obtain. To address these limitations, we propose an active data augmentation framework that synthesizes conversational training data by leveraging black-box LLMs guided by active learning techniques. Specifically, our method utilizes publicly available non-conversational domain data, including item metadata, user reviews, and collaborative signals, as seed inputs. By employing active learning strategies to select the most informative seed samples, our approach efficiently guides LLMs to generate synthetic, semantically coherent conversational interactions tailored explicitly to the target domain. Extensive experiments validate that conversational data generated by our proposed framework significantly improves the performance of LLM-based CRS models, effectively addressing the challenges of building CRS in no- or low-resource scenarios.

Figures

Figures reproduced from arXiv: 2504.15476 by the authors.

Figure 1
Figure 1. Comparison of CRS under different settings. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of our proposed Active Data Augmentation framework for conversational recommendation. Our pipeline [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Performance metrics for LlaMA3-1B across Redial [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: Performance metrics for LlaMA3-3B across Redial [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Performance metrics for Gemma2-2B across Redial [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: RQ3: Performance metrics (Recall@5 and NDCG@5) across iterations for different active learning strategies. [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: RQ4: Performance metrics (Recall@5 and NDCG@5) across iterations comparing collaborative-aware and semantic [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Comparison of model performance when trained [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

96 extracted references · 26 canonical work pages

  1. [1]

    Lora Aroyo, Alex Taylor, Mark Diaz, Christopher Homan, Alicia Parrish, Gre- gory Serapio-García, Vinodkumar Prabhakaran, and Ding Wang. 2023. Dices dataset: Diversity in conversational ai evaluation for safety. Advances in Neural Information Processing Systems 36 (2023), 53330–53342

    2023

  2. [2]

    Jordan Ash, Surbhi Goel, Akshay Krishnamurthy, and Sham Kakade. 2021. Gone fishing: Neural active learning with fisher embeddings. Advances in Neural Information Processing Systems 34 (2021), 8927–8939

    2021

  3. [3]

    Dong-Kyu Chae, Jihoo Kim, Duen Horng Chau, and Sang-Wook Kim. 2020. AR- CF: Augmenting virtual users and items in collaborative filtering for addressing cold-start problems. In Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval . 1251–1260

    2020

  4. [4]

    Lingjiao Chen, Matei Zaharia, and James Zou. 2023. Frugalgpt: How to use large language models while reducing cost and improving performance. arXiv preprint arXiv:2305.05176 (2023)

    arXiv 2023

  5. [5]

    Maximillian Chen, Alexandros Papangelis, Chenyang Tao, Seokhwan Kim, Andy Rosenbaum, Yang Liu, Zhou Yu, and Dilek Hakkani-Tur. 2023. PLACES: From Reviews to Dialogues: Active Synthesis for Zero-Shot LLM-based Conversational Recommender System Conference acronym ’XX, June 03–05, 2018, Woodstock, NY ReD INS Synth R+O INS+O 0 5 10 15 20Recall (%) Recall@1 on...

    arXiv 2023

  6. [6]

    Qibin Chen, Junyang Lin, Yichang Zhang, Ming Ding, Yukuo Cen, Hongxia Yang, and Jie Tang. 2019. Towards Knowledge-Based Recommender Dialog System. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Pro- cessing (EMNLP-IJCNLP), Kentaro Inui, Jing Jiang, Vi...

    doi:10.18653/v1/d19-1189 2019

  7. [7]

    Xiang Chen, Chaoyang Gao, Chunyang Chen, Guangbei Zhang, and Yong Liu

  8. [8]

    Zixiang Chen, Yihe Deng, Huizhuo Yuan, Kaixuan Ji, and Quanquan Gu. 2024. Self-play fine-tuning converts weak language models to strong language models. arXiv preprint arXiv:2401.01335 (2024)

    arXiv 2024

Show all 96 references
  1. [9]

    Konstantina Christakopoulou, Filip Radlinski, and Katja Hofmann. 2016. Towards conversational recommender systems. In Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining . 815–824

    2016

  2. [10]

    Flavio Di Palo, Prateek Singhi, and Bilal Fadlallah. 2024. Performance-Guided LLM Knowledge Distillation for Efficient Text Classification at Scale. arXiv preprint arXiv:2411.05045 (2024)

    2024 arXiv

  3. [11]

    Guanting Dong, Hongyi Yuan, Keming Lu, Chengpeng Li, Mingfeng Xue, Dayi- heng Liu, Wei Wang, Zheng Yuan, Chang Zhou, and Jingren Zhou. 2023. How abilities in large language models are affected by supervised fine-tuning data composition. arXiv preprint arXiv:2310.05492 (2023)

    2023 arXiv

  4. [12]

    Yingpeng Du, Zhu Sun, Ziyan Wang, Haoyan Chua, Jie Zhang, and Yew-Soon Ong. 2024. Active Large Language Model-based Knowledge Distillation for Session-based Recommendation. arXiv preprint arXiv:2502.15685 (2024)

    2024 arXiv

  5. [13]

    Yue Feng, Shuchang Liu, Zhenghai Xue, Qingpeng Cai, Lantao Hu, Peng Jiang, Kun Gai, and Fei Sun. 2023. A large language model enhanced conversational recommender system. arXiv preprint arXiv:2308.06212 (2023)

    2023 arXiv

  6. [14]

    R. A. Fisher. 1920. A mathematical Examination of the Methods of determin- ing the Accuracy of Observation by the Mean Error, and by the Mean Square Error. Monthly Notices of the Royal Astronomical Society 80, 8 (06 1920), 758–

    1920

  7. [15]

    Luke Friedman, Sameer Ahuja, David Allen, Zhenning Tan, Hakim Sidahmed, Changbo Long, Jun Xie, Gabriel Schubiner, Ajay Patel, Harsh Lara, et al. 2023. Leveraging large language models in conversational recommender systems.arXiv preprint arXiv:2305.07961 (2023)

    2023 arXiv

  8. [16]

    Drew Fudenberg, Gleb Romanyuk, and Philipp Strack. 2017. Active learning with a misspecified prior. Theoretical Economics 12, 3 (2017), 1155–1189

    2017

  9. [17]

    Silin Gao, Yichi Zhang, Zhijian Ou, and Zhou Yu. 2020. Paraphrase augmented task-oriented dialog generation. arXiv preprint arXiv:2004.07462 (2020)

    2020 arXiv

  10. [18]

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

    2024 arXiv

  11. [19]

    Shirley Anugrah Hayati, Dongyeop Kang, Qingxiaoyang Zhu, Weiyan Shi, and Zhou Yu. 2020. Inspired: Toward sociable recommendation dialog systems. arXiv preprint arXiv:2009.14306 (2020)

    2020 arXiv

  12. [20]

    Zhankui He, Zhouhang Xie, Rahul Jha, Harald Steck, Dawen Liang, Yesu Feng, Bodhisattwa Prasad Majumder, Nathan Kallus, and Julian McAuley. 2023. Large language models as zero-shot conversational recommenders. InProceedings of the 32nd ACM international conference on informatio...

    2023

  13. [21]

    Zhankui He, Handong Zhao, Tong Yu, Sungchul Kim, Fan Du, and Julian McAuley

  14. [22]

    Yupeng Hou, Jiacheng Li, Zhankui He, An Yan, Xiusi Chen, and Julian McAuley

  15. [23]

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, Weizhu Chen, et al. 2022. Lora: Low-rank adaptation of large language models. ICLR 1, 2 (2022), 3

    2022

  16. [24]

    Aaron Hurst, Adam Lerer, Adam P Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radford, et al. 2024. Gpt-4o system card. arXiv preprint arXiv:2410.21276 (2024)

    2024 arXiv

  17. [25]

    Ikotun, Absalom E

    Abiodun M. Ikotun, Absalom E. Ezugwu, Laith Abualigah, Belal Abuhaija, and Jia Heming. 2023. K-means clustering algorithms: A comprehensive review, variants analysis, and advances in the era of big data. Information Sciences 622 (2023), 178–210. doi:10.1016/j.ins.2022.11.139

    2023 doi

  18. [26]

    Pegah Jandaghi, Xianghai Sheng, Xinyi Bai, Jay Pujara, and Hakim Sidahmed

  19. [27]

    Cheonsu Jeong. 2024. Fine-tuning and utilization methods of domain-specific llms. arXiv preprint arXiv:2401.02981 (2024)

    2024 arXiv

  20. [28]

    Ting Jiang, Shaohan Huang, Zhongzhi Luan, Deqing Wang, and Fuzhen Zhuang

  21. [29]

    Enkelejda Kasneci, Kathrin Seßler, Stefan Küchemann, Maria Bannert, Daryna Dementieva, Frank Fischer, Urs Gasser, Georg Groh, Stephan Günnemann, Eyke Hüllermeier, et al. 2023. ChatGPT for good? On opportunities and challenges of large language models for education. Learning an...

    2023

  22. [30]

    Branislav Kveton, Xintong Li, Julian McAuley, Ryan Rossi, Jingbo Shang, Junda Wu, and Tong Yu. 2025. Active Learning for Direct Preference Optimization. arXiv preprint arXiv:2503.01076 (2025)

    2025 arXiv

  23. [31]

    In Findings of the Association for Computational Linguistics ACL 2024

    Faithful Persona-based Conversational Dataset Generation with Large Language Models. In Findings of the Association for Computational Linguistics ACL 2024. 15245–15270

    2024

  24. [32]

    Megan Leszczynski, Shu Zhang, Ravi Ganti, Krisztian Balog, Filip Radlinski, Fernando Pereira, and Arun Tejasvi Chaganty. 2023. Talk the walk: synthetic data generation for conversational music recommendation. arXiv preprint arXiv:2301.11489 (2023)

    2023 arXiv

  25. [33]

    Jiaxiang Li, Siliang Zeng, Hoi-To Wai, Chenliang Li, Alfredo Garcia, and Mingyi Hong. 2024. Getting more juice out of the sft data: Reward learning from human demonstration improves sft for llm alignment. Advances in Neural Information Processing Systems 37 (2024), 124292–124318

    2024

  26. [34]

    Qinbin Li, Junyuan Hong, Chulin Xie, Jeffrey Tan, Rachel Xin, Junyi Hou, Xavier Yin, Zhun Wang, Dan Hendrycks, Zhangyang Wang, et al. 2024. Llm-pbe: As- sessing data privacy in large language models. arXiv preprint arXiv:2408.12787 (2024)

    2024 arXiv

  27. [35]

    Quan Li, Tianxiang Zhao, Lingwei Chen, Junjie Xu, and Suhang Wang. 2024. En- hancing Graph Neural Networks with Limited Labeled Data by Actively Distilling Knowledge from Large Language Models. In 2024 IEEE International Conference on Big Data (BigData) . IEEE, 741–746

    2024

  28. [36]

    Raymond Li, Samira Ebrahimi Kahou, Hannes Schulz, Vincent Michalski, Laurent Charlin, and Chris Pal. 2018. Towards deep conversational recommendations. Advances in neural information processing systems 31 (2018)

    2018

  29. [37]

    Wenqiang Lei, Xiangnan He, Yisong Miao, Qingyun Wu, Richang Hong, Min- Yen Kan, and Tat-Seng Chua. 2020. Estimation-action-reflection: Towards deep interaction between conversational and recommender systems. In Proceedings of the 13th international conference on web search and...

    2020

  30. [38]

    Allen Lin, Ziwei Zhu, Jianling Wang, and James Caverlee. 2023. Enhancing user personalization in conversational recommenders. In Proceedings of the ACM Web Conference 2023. 770–778

    2023

  31. [39]

    Jianhua Lin. 1991. Divergence measures based on the Shannon entropy. IEEE Transactions on Information theory 37, 1 (1991), 145–151

    1991

  32. [40]

    Yong Lin, Chen Liu, Chenlu Ye, Qing Lian, Yuan Yao, and Tong Zhang. 2023. Optimal sample selection through uncertainty estimation and its application in deep learning. arXiv preprint arXiv:2309.02476 (2023). Conference acronym ’XX, June 03–05, 2018, Woodstock, NY Trovato et al

    2023 arXiv

  33. [41]

    Chengyuan Liu, Yangyang Kang, Fubang Zhao, Kun Kuang, Zhuoren Jiang, Changlong Sun, and Fei Wu. 2024. Evolving knowledge distillation with large language models and active learning. arXiv preprint arXiv:2403.06414 (2024)

    2024 arXiv

  34. [42]

    A Woodbury Max. 1950. Inverting modified matrices. In Memorandum Rept. 42, Statistical Research Group. Princeton Univ., 4

    1950

  35. [43]

    Ziyue Li and Tianyi Zhou. 2024. Your mixture-of-experts llm is secretly an embedding model for free. arXiv preprint arXiv:2410.10814 (2024)

    2024 arXiv

  36. [44]

    Subhojyoti Mukherjee, Anusha Lalitha, Kousha Kalantari, Aniket Anand Desh- mukh, Ge Liu, Yifei Ma, and Branislav Kveton. 2024. Optimal design for human preference elicitation. Advances in Neural Information Processing Systems 37 (2024), 90132–90159

    2024

  37. [45]

    Sheshera Mysore, Andrew McCallum, and Hamed Zamani. 2023. Large language model augmented narrative driven recommendations. In Proceedings of the 17th ACM Conference on Recommender Systems . 777–783

    2023

  38. [46]

    Seth Neel and Peter Chang. 2023. Privacy issues in large language models: A survey. arXiv preprint arXiv:2312.06717 (2023)

    2023 arXiv

  39. [47]

    Zooey Nguyen, Anthony Annunziata, Vinh Luong, Sang Dinh, Quynh Le, Anh Hai Ha, Chanh Le, Hong An Phan, Shruti Raghavan, and Christopher Nguyen. 2024. Enhancing Q&A with Domain-Specific Fine-Tuning and Iterative Reasoning: A Comparative Study. arXiv preprint arXiv:2404.11792 (2024)

    2024 arXiv

  40. [48]

    Yotam Perlitz, Ariel Gera, Michal Shmueli-Scheuer, Dafna Sheinwald, Noam Slonim, and Liat Ein-Dor. 2023. Active learning for natural language generation. arXiv preprint arXiv:2305.15040 (2023)

    2023 arXiv

  41. [49]

    Ahmed Menshawy, Zeeshan Nawaz, and Mahmoud Fahmy. 2024. Navigating Challenges and Technical Debt in Large Language Models Deployment. In Pro- ceedings of the 4th Workshop on Machine Learning and Systems . 192–199

    2024

  42. [50]

    Reddit. 2023. An Update Regarding Reddit’s API. https://www.reddit.com/r/reddit/ comments/12qwagm/an_update_regarding_reddits_api/ Accessed: 2025-03-30

    2023

  43. [51]

    Reddit. 2024. Publishing Our Public Content Policy and Introducing a New Commu- nity for Researchers. https://redditinc.com/blog/publishing-our-public-content- policy-and-introducing-a-new-community-for-researchers Accessed: 2025-03- 30

    2024

  44. [52]

    Xuhui Ren, Tong Chen, Quoc Viet Hung Nguyen, Lizhen Cui, Zi Huang, and Hongzhi Yin. 2024. Explicit knowledge graph reasoning for conversational recommendation. ACM Transactions on Intelligent Systems and Technology 15, 4 (2024), 1–21

    2024

  45. [53]

    Daniel C Ruiz and John Sell. 2024. Fine-Tuning and Evaluating Open-Source Large Language Models for the Army Domain. arXiv preprint arXiv:2410.20297 (2024)

    2024 arXiv

  46. [54]

    Siddharth Samsi, Dan Zhao, Joseph McDonald, Baolin Li, Adam Michaleas, Michael Jones, William Bergeron, Jeremy Kepner, Devesh Tiwari, and Vijay Gadepally. 2023. From words to watts: Benchmarking the energy costs of large language model inference. In 2023 IEEE High Performance ...

    2023

  47. [55]

    Reddit. 2023. Reddit Data API Update: Changes to Pushshift Access . https://www.reddit.com/r/modnews/comments/134tjpe/reddit_data_api_ update_changes_to_pushshift_access/ Accessed: 2025-03-30

    2023

  48. [56]

    Heydar Soudani, Roxana Petcu, Evangelos Kanoulas, and Faegheh Hasibi. 2024. A survey on recent advances in conversational data generation. arXiv preprint arXiv:2405.13003 (2024)

    2024 arXiv

  49. [57]

    Masashi Sugiyama. 2005. Active learning for misspecified models. Advances in neural information processing systems 18 (2005)

    2005

  50. [58]

    Jianwei Sun, Chaoyang Mei, Linlin Wei, Kaiyu Zheng, Na Liu, Ming Cui, and Tianyi Li. 2024. Dial-insight: Fine-tuning large language models with high- quality domain-specific data preventing capability collapse. arXiv preprint arXiv:2403.09167 (2024)

    2024 arXiv

  51. [59]

    Yueming Sun and Yi Zhang. 2018. Conversational recommender system. In The 41st international acm sigir conference on research & development in information retrieval. 235–244

    2018

  52. [60]

    Yixuan Tang and Yi Yang. 2024. Pooling and attention: What are effective designs for llm-based embedding models? arXiv preprint arXiv:2409.02727 (2024)

    2024 arXiv

  53. [61]

    Chandan Singh, Jeevana Priya Inala, Michel Galley, Rich Caruana, and Jianfeng Gao. 2024. Rethinking interpretability in the era of large language models. arXiv preprint arXiv:2402.01761 (2024)

    2024 arXiv

  54. [62]

    Tomasz Walkowiak and Mateusz Gniewkowski. 2019. Evaluation of vector em- bedding models in clustering of text documents. InProceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP 2019) . 1304–1311

    2019

  55. [63]

    Jian Wang, Yi Cheng, Dongding Lin, Chak Tou Leong, and Wenjie Li. 2023. Target- oriented proactive dialogue systems with personalization: Problem formulation and dataset curation. arXiv preprint arXiv:2310.07397 (2023)

    2023 arXiv

  56. [64]

    Xi Wang, Hossein A Rahmani, Jiqun Liu, and Emine Yilmaz. [n. d.]. Improving Conversational Recommendation Systems via Bias Analysis and Language-Model- Enhanced Data Augmentation. In The 2023 Conference on Empirical Methods in Natural Language Processing

    2023

  57. [65]

    Xiaolei Wang, Kun Zhou, Xinyu Tang, Wayne Xin Zhao, Fan Pan, Zhao Cao, and Ji-Rong Wen. 2023. Improving conversational recommendation systems via counterfactual data simulation. InProceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining . 2398–2408

    2023

  58. [66]

    Xiaolei Wang, Kun Zhou, Ji-Rong Wen, and Wayne Xin Zhao. 2022. Towards Unified Conversational Recommender Systems via Knowledge-Enhanced Prompt Learning. In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (Washington DC, USA) (KDD ’22). Ass...

    2022 doi

  59. [67]

    Gemma Team, Morgane Riviere, Shreya Pathak, Pier Giuseppe Sessa, Cassidy Hardin, Surya Bhupatiraju, Léonard Hussenot, Thomas Mesnard, Bobak Shahriari, Alexandre Ramé, et al. 2024. Gemma 2: Improving open language models at a practical size. arXiv preprint arXiv:2408.00118 (2024)

    2024 arXiv

  60. [68]

    Qingyang Wu, Song Feng, Derek Chen, Sachindra Joshi, Luis A Lastras, and Zhou Yu. 2021. DG2: Data augmentation through document grounded dialogue generation. arXiv preprint arXiv:2112.08342 (2021)

    2021 arXiv

  61. [69]

    Yunjia Xi, Weiwen Liu, Jianghao Lin, Bo Chen, Ruiming Tang, Weinan Zhang, and Yong Yu. 2024. MemoCRS: Memory-enhanced Sequential Conversational Recommender Systems with Large Language Models. In Proceedings of the 33rd ACM International Conference on Information and Knowledge ...

    2024

  62. [70]

    Yu Xia, Xu Liu, Tong Yu, Sungchul Kim, Ryan A Rossi, Anup Rao, Tung Mai, and Shuai Li. 2024. Hallucination diversity-aware active learning for text summariza- tion. arXiv preprint arXiv:2404.01588 (2024)

    2024 arXiv

  63. [71]

    Yu Xia, Subhojyoti Mukherjee, Zhouhang Xie, Junda Wu, Xintong Li, Ryan Aponte, Hanjia Lyu, Joe Barrow, Hongjie Chen, Franck Dernoncourt, et al. 2025. From Selection to Generation: A Survey of LLM-based Active Learning. arXiv preprint arXiv:2502.11767 (2025)

    2025 arXiv

  64. [72]

    Yu Xia, Junda Wu, Tong Yu, Sungchul Kim, Ryan A Rossi, and Shuai Li. 2023. User-regulation deconfounded conversational recommender system with ban- dit feedback. In Proceedings of the 29th ACM SIGKDD conference on knowledge discovery and data mining . 2694–2704

    2023

  65. [73]

    Likang Wu, Zhi Zheng, Zhaopeng Qiu, Hao Wang, Hongchao Gu, Tingjia Shen, Chuan Qin, Chen Zhu, Hengshu Zhu, Qi Liu, et al . 2024. A survey on large language models for recommendation. World Wide Web 27, 5 (2024), 60

    2024

  66. [74]

    Biwei Yan, Kun Li, Minghui Xu, Yueyan Dong, Yue Zhang, Zhaochun Ren, and Xiuzhen Cheng. 2024. On protecting the data privacy of large language models (llms): A survey. arXiv preprint arXiv:2403.05156 (2024)

    2024 arXiv

  67. [75]

    Fan Yang, Zheng Chen, Ziyan Jiang, Eunah Cho, Xiaojiang Huang, and Yanbin Lu. 2023. Palr: Personalization aware llms for recommendation. arXiv preprint arXiv:2305.07622 (2023)

    2023 arXiv

  68. [76]

    Ting Yang and Li Chen. 2024. Unleashing the Retrieval Potential of Large Lan- guage Models in Conversational Recommender Systems. In Proceedings of the 18th ACM Conference on Recommender Systems . 43–52

    2024

  69. [77]

    Yifan Yao, Jinhao Duan, Kaidi Xu, Yuanfang Cai, Zhibo Sun, and Yue Zhang. 2024. A survey on large language model (llm) security and privacy: The good, the bad, and the ugly. High-Confidence Computing (2024), 100211

    2024

  70. [78]

    Wei Jie Yeo, Ranjan Satapathy, Rick Siow Mong Goh, and Erik Cambria. 2024. How interpretable are reasoning explanations from prompting large language models? arXiv preprint arXiv:2402.11863 (2024)

    2024 arXiv

  71. [79]

    Zhouhang Xie, Junda Wu, Hyunsik Jeon, Zhankui He, Harald Steck, Rahul Jha, Dawen Liang, Nathan Kallus, and Julian McAuley. 2024. Neighborhood-Based Collaborative Filtering for Conversational Recommendation. In Proceedings of the 18th ACM Conference on Recommender Systems . 1045–1050

    2024

  72. [80]

    Shengyu Zhang, Linfeng Dong, Xiaoya Li, Sen Zhang, Xiaofei Sun, Shuhe Wang, Jiwei Li, Runyi Hu, Tianwei Zhang, Fei Wu, et al. 2023. Instruction tuning for large language models: A survey. arXiv preprint arXiv:2308.10792 (2023)

    2023

  73. [81]

    Yongfeng Zhang, Xu Chen, Qingyao Ai, Liu Yang, and W Bruce Croft. 2018. Towards conversational search and recommendation: System ask, user respond. In Proceedings of the 27th acm international conference on information and knowledge management. 177–186

    2018

  74. [82]

    Yifei Zhang, Bo Pan, Chen Ling, Yuntong Hu, and Liang Zhao. 2024. ELAD: Explanation-guided large language models active distillation. arXiv preprint arXiv:2402.13098 (2024)

    2024 arXiv

  75. [83]

    Yiming Zhang, Lingfei Wu, Qi Shen, Yitong Pang, Zhihua Wei, Fangli Xu, Bo Long, and Jian Pei. 2022. Multiple choice questions based multi-interest policy learning for conversational recommendation. In Proceedings of the ACM Web Conference 2022. 2153–2162

    2022

  76. [84]

    Zhipeng Zhao, Kun Zhou, Xiaolei Wang, Wayne Xin Zhao, Fan Pan, Zhao Cao, and Ji-Rong Wen. 2023. Alleviating the long-tail problem in conversational recommender systems. InProceedings of the 17th ACM Conference on Recommender Systems. 374–385

    2023

  77. [85]

    Junjie Zhang, Ruobing Xie, Yupeng Hou, Xin Zhao, Leyu Lin, and Ji-Rong Wen

  78. [86]

    ACM Transactions on Information Systems (2023)

    Recommendation as instruction following: A large language model em- powered recommendation approach. ACM Transactions on Information Systems (2023)

    2023

  79. [87]

    Jinfeng Zhou, Bo Wang, Ruifang He, and Yuexian Hou. 2021. CRFR: Improving conversational recommender systems via flexible fragments reasoning on knowl- edge graphs. In Proceedings of the 2021 conference on empirical methods in natural language processing. 4324–4334

    2021

  80. [88]

    Kun Zhou, Wayne Xin Zhao, Shuqing Bian, Yuanhang Zhou, Ji-Rong Wen, and Jingsong Yu. 2020. Improving conversational recommender systems via knowl- edge graph based semantic fusion. In Proceedings of the 26th ACM SIGKDD international conference on knowledge discovery & data min...

    2020

  81. [92]

    Zhi Zheng, Wenshuo Chao, Zhaopeng Qiu, Hengshu Zhu, and Hui Xiong. 2024. Harnessing large language models for text-rich sequential recommendation. In Proceedings of the ACM Web Conference 2024 . 3207–3216

    2024

  82. [93]

    Zhi Zheng, Zhaopeng Qiu, Xiao Hu, Likang Wu, Hengshu Zhu, and Hui Xiong

  83. [94]

    Generative job recommendations with large language model.arXiv preprint From Reviews to Dialogues: Active Synthesis for Zero-Shot LLM-based Conversational Recommender System Conference acronym ’XX, June 03–05, 2018, Woodstock, NY arXiv:2307.02157 (2023)

    2023 arXiv

  84. [770]

    doi:10.1093/mnras/80.8.758 arXiv:https://academic.oup.com/mnras/article- pdf/80/8/758/3655524/mnras80-0758.pdf

    doi

  85. [2022]

    In Proceed- ings of the 16th ACM Conference on Recommender Systems

    Bundle MCR: Towards conversational bundle recommendation. In Proceed- ings of the 16th ACM Conference on Recommender Systems . 288–298

  86. [2023]

    arXiv preprint arXiv:2307.16645 (2023)

    Scaling sentence embeddings with large language models. arXiv preprint arXiv:2307.16645 (2023)

    2023 arXiv

  87. [2024]

    arXiv preprint arXiv:2403.03952 (2024)

    Bridging language and items for retrieval and recommendation. arXiv preprint arXiv:2403.03952 (2024)

    2024 arXiv

  88. [2025]

    ACM Transactions on Software Engineering and Methodology (2025)

    An Empirical Study on Challenges for LLM Application Developers. ACM Transactions on Software Engineering and Methodology (2025)

    2025

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.