Pith. sign in

REVIEW 4 major objections 4 minor 2 cited by

Dynamic Knowledge Exchange and Dual-diversity Review: Concisely Unleashing the Potential of a Multi-Agent Research Team

T0 review · 4 major / 4 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 A multi-agent LLM team with dynamic knowledge exchange and dual-diversity review generates more novel and impactful scientific abstracts than prior LLM scientist systems, in both computer science and a new health-sciences benchmark.

desk verdict A useful new dataset and a coherent multi-agent pipeline, but the evaluation metrics are circular proxies and the paper's own example undermines the central outperformance claim. read the letter →

arxiv 2506.18348 v3 pith:6WLRNZZY submitted 2025-06-23 cs.AI

Weilun Yu , Shixiang Tang , Yonggui Huang , Nanqing Dong , Li Fan , Honggang Qi , Wei Liu , Xiaoli Diao
show 2 more authors
Xi Chen Wanli Ouyang
This is my paper · ORCID
classification cs.AI
keywords multi-agentLLMsystemsscientificideagenerationdynamicknowledgeexchangedual-diversityreviewautomatednoveltymetricshealthsciencesdatasetBordacountvotingLLM-basedresearchagents
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 sets out to show that team structure, not just model size or agent count, determines the quality of ideas an LLM research team produces. It proposes IDVSCI (Internal Discussion and Vote SCIentists), built on two mechanisms: agents revise one another's ideas through a leader-mediated exchange, and reviewers with heterogeneous backgrounds score ideas using literature-grounded, dynamically updated prompts. On four automated novelty-and-impact metrics, IDVSCI reports better scores than two prior LLM scientist pipelines in both computer science and a new health-sciences dataset, with the largest tested model reaching a Contemporary Impact of 10.06 and an Overall Novelty of 10.58 in health sciences. If these proxy metrics track real scientific value, the result would support designing autonomous research teams around structured debate and diverse review rather than fixed sequential role-playing.

What carries the argument

Dynamic Knowledge Exchange reorganizes idea generation around a leader who aggregates cross-agent revisions: each scientist drafts an idea, other scientists revise it from their own knowledge bases, and the leader synthesizes the revisions before the originator reflects and finalizes. Dual-Diversity Review makes the team heterogeneous in background and refreshes each agent's prompt with the top-k most relevant retrieved references, then ranks candidate ideas by a confidence-weighted Borda count. The load-bearing mechanism is the combination of internal cross-revision with diverse, literature-grounded evaluation, which the paper argues prevents repetitive convergence and surfaces more original ideas.

What would settle it

Ask domain experts, blind to which system produced each abstract, to rate novelty and feasibility; if their rankings do not reproduce the Contemporary Impact and Overall Novelty ordering, or if re-running with a different embedding model reverses the ranking, the reported advantage is an artifact of the metric.

Watch

Extended reading notes

Core claim

IDVSCI's central claim is that Dynamic Knowledge Exchange and Dual-Diversity Review transform a group of LLM agents into a research team whose generated abstracts are more dissimilar from pre-2011 work, more aligned with recent highly cited work, and more likely to resemble impactful research. In the computer-science benchmark the best configuration raises Contemporary Impact and Overall Novelty above all baselines while retaining comparable dissimilarity scores; in the new health-sciences benchmark it raises Contemporary Impact to 10.06 and Overall Novelty to 10.58. The paper frames these results as consistent outperformance relative to prior systems across both domains.

Load-bearing premise

The entire comparison rests on treating four automated scores—distance from old abstracts, closeness to recent abstracts, and citation counts of similar recent abstracts—as proxies for how novel and impactful a generated idea would be in real science.

Editorial extensions

If this is right

  • The paper reports that IDVSCI outperforms both prior LLM scientist pipelines on the four metrics in both domains, with the largest gains in Contemporary Impact and Overall Novelty.
  • The largest tested configuration reaches the strongest metrics in health sciences, with Contemporary Impact 10.06 and Overall Novelty 10.58.
  • Removing the internal discussion module hurts performance most, dropping Contemporary Impact from 4.38 to 4.10 in the computer-science setup.
  • One or two discussion rounds are nearly enough: further rounds add little and can even lower scores, so the framework is cheap to run.
  • The voting step can slightly cap impact but protects originality, since removing it raises Contemporary Impact while lowering Overall Novelty.

Reading between the lines

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

  • Editorial: the four metrics reward abstracts that sit close to recent highly cited work, so part of the measured gain may reflect convergence on popular research themes rather than genuinely new directions.
  • Editorial: the strongest test of the claim is human expert review; if blind domain experts do not rank IDVSCI abstracts above the baselines, the reported advantage is metric-specific.
  • Editorial: re-running the pipeline with a different embedding model would show whether the HD/CD/CI/ON ranking is stable or an artifact of the chosen distance measure.
  • Editorial: the health-sciences examples cluster around gratitude-mindfulness interventions, so measuring topical diversity across repeated runs could separate breadth from the metric scores.
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 / 4 minor

Summary. The paper proposes IDVSCI, an LLM-based multi-agent framework for generating scientific abstracts, with two main components: Dynamic Knowledge Exchange (DKE), where agents revise each other's ideas and a leader synthesizes feedback, and Dual-Diversity Review (DDR), which diversifies agent teams and prompts and uses a weighted Borda-count vote to select ideas. The authors compare IDVSCI against AI-Scientist and VIRSCI on a computer-science dataset and on a newly introduced health-sciences dataset, using four embedding- and citation-based metrics: Historical Dissimilarity (HD), Contemporary Dissimilarity (CD), Contemporary Impact (CI), and Overall Novelty (ON). The paper reports that IDVSCI achieves the best performance on both datasets and includes ablations on module removal, iteration count, background diversity, team size, and internal versus external review.

Significance. If the mechanisms work as described, this would be a modest systems contribution: DKE and DDR are plausible ways to structure multi-agent ideation, and the authors provide code, both datasets, and a two-domain test with 20 trials per condition. The ablations in Tables 2-4 and Figures 4-5 target the main design choices. However, the significance cannot be assessed from the reported evidence because the evaluation rests on unvalidated proxy metrics, the headline claim is inconsistent with Table 1, and the only displayed example of the system's output is internally contradictory. The contribution is better characterized as a framework description than as a demonstration of improved scientific discovery.

major comments (4)
  1. [Results, Table 1] The claim in the Results section that IDVSCI "consistently outperforms AI-Scientist and VIRSCI across all metrics" is contradicted by Table 1. In Computer Sciences, the three IDVSCI rows have HD values 0.40, 0.41, and 0.40, which are lower than the AI-Scientist values 0.51 and 0.53 and lower than the VIRSCI values 0.43, 0.46, and 0.44, despite HD being a higher-is-better metric. In Health Sciences, every IDVSCI row has a CD value (0.39, 0.40, 0.39) that is worse than the corresponding VIRSCI values (0.37, 0.38, 0.37), on a lower-is-better metric. The central empirical claim therefore needs to be restated and re-supported with a comparison that respects the direction of each metric.
  2. [Evaluation Metrics; Figure 7] The four proxy metrics are not validated as measures of scientific novelty or impact, and the paper's own example shows the failure mode. In Figure 7, the IDVSCI (LLaMA-70b) abstract describes a randomized controlled trial in which participants "will be recruited" and "will participate," yet later asserts "Preliminary findings indicate significant improvements" and "fMRI scans reveal" results, an internal contradiction that a scientifically sound abstract cannot contain. The same abstract receives CI = 17.63. Because CI is defined as the citation count of the five nearest post-2011 neighbors, a derivative abstract that paraphrases citation-rich material can score highly, while a genuinely original idea with no close neighbors would score low. No human evaluation, feasibility check, or factual-soundness validation is reported, so Table 1 does not establish that IDVSCI produces better scientific ideas.
  3. [Experimental Setup; Table 1] The manuscript states that all metrics are averages over 20 independent trials, but Table 1 reports no standard deviation, confidence interval, or significance test. Many of the reported differences are small (e.g., CI 4.38 versus 4.17 for IDVSCI with LLaMA-8b versus QWQ-32b in Computer Sciences), and the ablation discussion describes changes as "significant" without uncertainty quantification. Without variance information, the claim of consistent superiority and the characterization of ablation effects cannot be evaluated.
  4. [Evaluation Metrics (ON)] The statement in the Evaluation Metrics section that "Mathematically, the expected value of ON is proportional to the true novelty" is a non-sequitur. ON is defined as HD x CI / CD, but no definition of "true novelty" is supplied, no derivation is given, and the claimed proportionality is not evident. Because ON is used as a headline metric, this assertion must be either proven, replaced with a defined quantity, or removed.
minor comments (4)
  1. [Abstract] The abstract contains a typographical artifact: "Internal Discussion and V ote SCIentists" should be "Vote."
  2. [Figure 6] The Figure 6 caption and surrounding text contain stray glyphs such as "🕮", "¶agic", and "♂" that appear to be encoding artifacts; these should be cleaned before submission.
  3. [Results, Health Sciences] All ablations and analyses after Table 1 are performed only on the Computer Sciences dataset, so the cross-domain adaptability claim for the Health Sciences dataset rests on a single table with no health-domain ablations or analysis.
  4. [Figures 7 and 8] The generated outputs in Figures 7 and 8 are multi-paragraph documents, whereas the corpus abstracts used as metric references are single paragraphs; this length mismatch may bias the embedding-distance comparisons and should be discussed or controlled.

Circularity Check

2 steps flagged · score 6.0 of 10

Evaluation is partly circular: CI/CD are computed from the same retrieval embeddings that shape the generated abstracts, and ON is asserted to measure 'true novelty' by definition.

  1. fitted input called prediction [Method, Dual-Diversity Review Eq. (3); Experiment, Evaluation Metrics]
    "we leverage the Faiss library (Douze et al. 2024) to compute the Euclidean distances between each generated idea and papers in the database. The top-k nearest papers are retrieved as reference material... P′_i = Pi∪{ Paper_j| Paper_j∈ Top-k(D(I, Paper))} ... Contemporary Impact (CI): This reflects the citation count of the top 5 most similar abstracts in the corpus of articles published after 2011."

    The generation prompts are updated with the top-k nearest corpus abstracts by Euclidean distance in exactly the embedding space used to compute the evaluation metrics. CI is then the citation count of the 5 nearest post-2011 abstracts, and CD is the distance to those same nearest neighbors. A generated abstract that paraphrases a retrieved, highly cited paper will be close to that paper, yielding high CI and low CD. Thus the headline 'impact' and 'contemporary dissimilarity' scores are not independent measurements of the abstract's own quality; they are k-NN citation lookups on the documents the system was prompted to emulate. The Table 1 comparison therefore partly measures how well the retrieval loop is exploited, not the scientific merit of the generated ideas.

  2. self definitional [Experiment, Evaluation Metrics (Overall Novelty definition)]
    "Overall Novelty (ON): ON is positively correlated with HD and CI and negatively correlated with CD. It is calculated as ON = HD×CI/CD. Mathematically, the expected value of ON is proportional to the true novelty."

    No independent definition of 'true novelty' is given anywhere in the paper. The assertion that the expected value of ON is proportional to the true novelty either defines 'true novelty' as what ON measures or is an unproved leap. The paper then uses ON as a headline indicator of innovativeness in Table 1, so the conclusion that IDVSCI is more novel because its ON is highest is a restatement of the metric rather than an empirically supported statement about an externally anchored quantity.

full rationale

The proposed DKE and DDR mechanisms are engineering contributions whose design is not itself derived from the evaluation metrics; there is no self-citation chain that forces the framework choices. The circularity is concentrated in the evaluation loop. Eq. (3) inserts the top-k nearest corpus abstracts (by Faiss Euclidean distance in the mxbai-embed-large space) into the agents' prompts, and the headline metrics HD, CD, and CI are computed from the 5 nearest abstracts in the same embedding space, with CI being the citation counts of those neighbors. Consequently, a high CI and low CD are partly consequences of the retrieval loop rather than independent evidence of the generated abstract's impact or novelty. The further claim that 'the expected value of ON is proportional to the true novelty' is a definitional assertion: no external quantity called 'true novelty' is defined, so the highest-ON result in Table 1 is a restatement of the metric. The metrics are introduced as following VIRSCI's methodology, a paper with overlapping authors, which is a self-citation, but the more direct concern is the shared retrieval/evaluation embedding space. The paper itself acknowledges that the metrics are a 'partial' evaluation, and no external human validation or factual-soundness check is reported, so the central superiority claim rests entirely on the proxy metrics. This is partial circularity rather than a fully forced derivation.

Assumptions & free parameters 4 free parameters · 4 assumptions · 2 invented entities

The paper rests on a small set of free parameters: the Top-k retrieval count, the number of discussion rounds, team size, and the diversity level, each chosen by hand or inherited from baselines. The core evaluation depends on the assumption that embedding-distance and citation proxies capture scientific novelty. The framework itself is an invented system evaluated only on its own metrics, while the new dataset is the only independently reusable artifact.

free parameters (4)
  • Top-k number of retrieved papers = 8 (check novelty stage); empirically set, as stated in 'Experimental Setup'
    The paper states 'we empirically set k = 8 for Top-k(D(I, Paperj)) in the Check Novelty phase'; the choice is justified by an observational study, not by a principled criterion.
  • Number of discussion rounds = 5 in main experiments; ablation shows 1-2 rounds are near-optimal
    The main results use five sequential rounds, following the VIRSCI protocol, but the iteration ablation shows no clear benefit beyond one or two rounds, indicating the choice is not independently justified.
  • Team size = 4 in main experiments; 8 in a secondary experiment
    Team size is inherited from VIRSCI's optimal configuration and is not optimized for IDVSCI; the paper admits no significant improvement when scaling to 8.
  • Background diversity level = 25% optimal in the diversity ablation
    The paper reports best performance at 25% diversity, but this level is discovered post hoc from the ablation and is not derived from theory.
assumptions (4)
  • domain assumption The proxy metrics HD, CD, CI, ON measure scientific novelty and impact.
    The entire comparison rests on embedding distance and citation-count proxies. The paper cites prior work using these metrics but does not validate them against human judgments or actual scientific merit.
  • domain assumption A single generated abstract is a sufficient unit for evaluating the scientific quality of an ideation system.
    The system is tasked to produce one abstract; the evaluation ignores whether the underlying ideas are feasible, correct, or reproducible.
  • domain assumption LLM agents with different knowledge bases and prompts reliably emulate distinct expert perspectives.
    The DDR mechanism assumes that varying prompts and retrieved papers yields heterogeneous expert evaluation; the paper's own observation that more references cause agents to produce identical outputs reveals the fragility of this assumption.
  • domain assumption Weighted Borda count with self-assigned confidence scores is a valid aggregation of expert judgments.
    The voting scheme uses LLM-assigned confidence scores as weights, but no evidence is given that these confidence values are calibrated.
invented entities (2)
  • IDVSCI framework (Internal Discussion and Vote SCIentists)
    purpose: Proposed multi-agent system for generating scientific abstracts with iterative feedback and dual-diversity voting.
    The framework is introduced as the paper's contribution; its performance is evaluated only on the paper's own proxy metrics, with no external validation of the scientific quality of its outputs.
  • Health Sciences Dataset independent evidence
    purpose: New evaluation dataset assembled from PubMed with 130 researchers and 86,448 publications.
    The dataset is a concrete artifact shared by the authors; it can be reused by other groups, providing a falsifiable handle outside the paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Dynamic Knowledge Exchange and Dual-diversity Review: Concisely Unleashing the Potential of a Multi-Agent Research Team." pith.science (2026). https://pith.science/paper/6WLRNZZY

@misc{pith2026250618348,
  author       = {Pith},
  title        = {Pith review of: Dynamic Knowledge Exchange and Dual-diversity Review: Concisely Unleashing the Potential of a Multi-Agent Research Team},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6WLRNZZY}},
  note         = {Machine review of arXiv:2506.18348}
}
read the original abstract

Scientific progress increasingly relies on effective collaboration among researchers, a dynamic that large language models (LLMs) have only begun to emulate. While recent LLM-based scientist agents show promise in autonomous scientific discovery, they often lack the interactive reasoning and evaluation mechanisms essential to real-world research. We propose IDVSCI (Internal Discussion and Vote SCIentists), a multi-agent framework built on LLMs that incorporates two key innovations: a Dynamic Knowledge Exchange mechanism enabling iterative feedback among agents, and a Dual-Diversity Review paradigm that simulates heterogeneous expert evaluation. These components jointly promote deeper reasoning and the generation of more creative and impactful scientific ideas. To evaluate the effectiveness and generalizability of our approach, we conduct experiments on two datasets: a widely used benchmark in computer science and a new dataset we introduce in the health sciences domain. Results show that IDVSCI consistently achieves the best performance across both datasets, outperforming existing systems such as AI Scientist and VIRSCI. These findings highlight the value of modeling interaction and peer review dynamics in LLM-based autonomous research.

Figures

Figures reproduced from arXiv: 2506.18348 by the authors.

Figure 1
Figure 1. Illustration of the workflow comparison between [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. The workflow framework of our method designed for automated scientific research. It consists of four sequential steps: [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. A case study illustrates the responses generated by [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (5 more)
Figure 5
Figure 5. Figure 5: Comparison chart of results when team size is 8. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 4
Figure 4. Figure 4: Comparative experimental data for visualizing [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 6
Figure 6. Figure 6: Health Sciences Dataset as the foundation for ecosystem construction and agent initialization. [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Example abstract generated by our IDVSCI model, focusing on psychological rehabilitation and intervention for [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 8
Figure 8. Figure 8: Example abstract generated by the baseline VIRSCI model, also targeting psychological rehabilitation and inter [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

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

  1. Diversifying Personalized Research Ideation against AI-Induced Homogenization

    cs.AI 2026-07 conditional novelty 5.5 of 10

    DivAlign cuts community nearest-neighbor idea similarity from 0.663 to 0.608 while keeping about 99.9% of researcher–direction fit versus independent top-choice selection.

  2. Matching Game Preferences Through Dialogical Large Language Models: A Perspective

    cs.AI 2025-07 conditional novelty 4.0 of 10

    This perspective paper proposes the D-LLM framework, which couples the authors' GRAPHYP knowledge graphs with LLMs to personalize AI responses and make reasoning traceable, but no empirical validation is presented.

Reference graph

Works this paper leans on

37 extracted references · 23 canonical work pages · cited by 2 Pith papers

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    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 word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    Agrawal, H.; Mishra, A.; Gupta, M.; et al. 2023. Multimodal Persona Based Generation of Comic Dialogs. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 14150--14164

    work page 2023

  4. [4]

    Chen, J.; Wang, X.; Xu, R.; Yuan, S.; Zhang, Y.; Shi, W.; Xie, J.; Li, S.; Yang, R.; Zhu, T.; Chen, A.; Li, N.; Chen, L.; Hu, C.; Wu, S.; Ren, S.; Fu, Z.; and Xiao, Y. 2024 a . From Persona to Personalization: A Survey on Role-Playing Language Agents. Transactions on Machine Learning Research. Survey Certification

    work page 2024

  5. [5]

    Chen, W.; Su, Y.; Zuo, J.; Yang, C.; Yuan, C.; Chan, C.-M.; Yu, H.; Lu, Y.; Hung, Y.-H.; Qian, C.; Qin, Y.; Cong, X.; Xie, R.; Liu, Z.; Sun, M.; and Zhou, J. 2024 b . AgentVerse: Facilitating Multi-Agent Collaboration and Exploring Emergent Behaviors. In The Twelfth International Conference on Learning Representations

    work page 2024

  6. [6]

    Douze, M.; Guzhva, A.; Deng, C.; Johnson, J.; Szilvasy, G.; Mazar \'e , P.-E.; Lomeli, M.; Hosseini, L.; and J \'e gou, H. 2024. The faiss library. arXiv preprint arXiv:2401.08281

    arXiv 2024

  7. [7]

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

    arXiv 2024

  8. [8]

    Gao, D.; Li, Z.; Pan, X.; Kuang, W.; Ma, Z.; Qian, B.; Wei, F.; Zhang, W.; Xie, Y.; Chen, D.; et al. 2024 a . Agentscope: A flexible yet robust multi-agent platform. arXiv preprint arXiv:2402.14034

    arXiv 2024

Show all 37 references
  1. [9]

    Gao, S.; Li, H.; Shi, Z.; Huang, C.; Tu, Q.; Shang, S.; Tian, Z.; and Huang, M. 2024 b . 360 ^ REA : Towards A Reusable Experience Accumulation with 360 Assessment for Multi-Agent System. In Findings of the Association for Computational Linguistics ACL 2024, 13149--13162

    2024

  2. [10]

    Gauch, H. G. 2003. Scientific method in practice. Cambridge University Press

    2003

  3. [11]

    Ghafarollahi, A.; and Buehler, M. 2024 a . ProtAgents: Protein discovery via large language model multi-agent collaborations combining physics and machine learning. In ICLR 2024 Workshop on Large Language Model (LLM) Agents

    2024

  4. [12]

    Ghafarollahi, A.; and Buehler, M. J. 2024 b . AtomAgents: Alloy design and discovery through physics-aware multi-modal multi-agent artificial intelligence. arXiv preprint arXiv:2407.10022

    2024 arXiv

  5. [13]

    Ghafarollahi, A.; and Buehler, M. J. 2024 c . SciAgents: Automating Scientific Discovery Through Bioinspired Multi-Agent Intelligent Graph Reasoning. Advanced Materials, 2413523

    2024

  6. [14]

    Guangyao, C.; Siwei, D.; Yu, S.; Ge, Z.; Jaward, S.; Börje, K.; Jie, F.; and Yemin, S. 2024. AutoAgents: A Framework for Automatic Agent Generation. In IJCAI, 22--30

    2024

  7. [15]

    Hong, S.; Zhuge, M.; Chen, J.; Zheng, X.; Cheng, Y.; Wang, J.; Zhang, C.; Wang, Z.; Yau, S. K. S.; Lin, Z.; Zhou, L.; Ran, C.; Xiao, L.; Wu, C.; and Schmidhuber, J. 2024. Meta GPT : Meta Programming for A Multi-Agent Collaborative Framework. In The Twelfth International Confer...

    2024

  8. [16]

    Lee, S.; Shakir, A.; Koenig, D.; and Lipp, J. 2024. Open source strikes bread-new fluffy embeddings model

    2024

  9. [17]

    Liu, W.; Wang, C.; Wang, Y.; Xie, Z.; Qiu, R.; Dang, Y.; Du, Z.; Chen, W.; Yang, C.; and Qian, C. 2024 a . Autonomous Agents for Collaborative Task under Information Asymmetry. In The Thirty-eighth Annual Conference on Neural Information Processing Systems

    2024

  10. [18]

    Liu, Y.; Shen, R.; Zhou, L.; Xiao, Q.; Yuan, J.; and Li, Y. 2024 b . Harnessing Data-Intelligence-Intensive Multi-Agent System for Life Science Research. bioRxiv, 2024--05

    2024

  11. [19]

    T.; Foerster, J.; Clune, J.; and Ha, D

    Lu, C.; Lu, C.; Lange, R. T.; Foerster, J.; Clune, J.; and Ha, D. 2024. The ai scientist: Towards fully automated open-ended scientific discovery. arXiv preprint arXiv:2408.06292

    2024 arXiv

  12. [20]

    Niu, R.; Li, J.; Wang, S.; Fu, Y.; Hu, X.; Leng, X.; Kong, H.; Chang, Y.; and Wang, Q. 2024. ScreenAgent: A Vision Language Model-driven Computer Control Agent. In Larson, K., ed., Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence, IJCAI...

    2024

  13. [21]

    OpenAI . 2022. Chatgpt. Technical report, OpenAI

    2022

  14. [22]

    Qian, C.; Liu, W.; Liu, H.; Chen, N.; Dang, Y.; Li, J.; Yang, C.; Chen, W.; Su, Y.; Cong, X.; Xu, J.; Li, D.; Liu, Z.; and Sun, M. 2024. C hat D ev: Communicative Agents for Software Development. In Ku, L.-W.; Martins, A.; and Srikumar, V., eds., Proceedings of the 62nd Annual...

    2024

  15. [23]

    Schmidgall, S.; Su, Y.; Wang, Z.; Sun, X.; Wu, J.; Yu, X.; Liu, J.; Liu, Z.; and Barsoum, E. 2025. Agent Laboratory: Using LLM Agents as Research Assistants. arXiv preprint arXiv:2501.04227

    2025 arXiv

  16. [24]

    Sclar, M.; Kumar, S.; West, P.; Suhr, A.; Choi, Y.; and Tsvetkov, Y. 2023. Minding Language Models’(Lack of) Theory of Mind: A Plug-and-Play Multi-Character Belief Tracker. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Lo...

    2023

  17. [25]

    Shanahan, M.; McDonell, K.; and Reynolds, L. 2023. Role play with large language models. Nature, 623(7987): 493--498

    2023

  18. [26]

    Shang, S.; Yao, Z.; Fu, H.; Tao, C.; Chen, X.; Wang, F.; Wang, Y.; Ren, Z.; and Gao, S. 2024. Unified Multi-Scenario Summarization Evaluation and Explanation. IEEE Transactions on Knowledge and Data Engineering

    2024

  19. [27]

    Shao, Y.; Mao, J.; Liu, Y.; Ma, W.; Satoh, K.; Zhang, M.; and Ma, S. 2020. BERT-PLI: Modeling paragraph-level interactions for legal case retrieval. In IJCAI, 3501--3507

    2020

  20. [28]

    Song, T.; Luo, M.; Chen, L.; Huang, Y.; Zhu, Q.; Liu, D.; Zhang, B.; Zou, G.; Zhang, F.; Shang, W.; et al. 2024. A multi-agent-driven robotic AI chemist enabling autonomous chemical research on demand. chemrxiv

    2024

  21. [29]

    Su, H.; Chen, R.; Tang, S.; Yin, Z.; Zheng, X.; Li, J.; Qi, B.; Wu, Q.; Li, H.; Ouyang, W.; Torr, P.; Zhou, B.; and Dong, N. 2025. Many Heads Are Better Than One: Improved Scientific Idea Generation by A LLM-Based Multi-Agent System. In Proceedings of the 63nd Annual Meeting o...

    2025

  22. [30]

    Tang, J.; Zhang, J.; Yao, L.; Li, J.; Zhang, L.; and Su, Z. 2008. Arnetminer: extraction and mining of academic social networks. In Proceedings of the 14th ACM SIGKDD international conference on Knowledge discovery and data mining, 990--998

    2008

  23. [31]

    Team, N.; Zhang, B.; Feng, S.; Yan, X.; Yuan, J.; Yu, Z.; He, X.; Huang, S.; Hou, S.; Nie, Z.; et al. 2025. NovelSeek: When Agent Becomes the Scientist--Building Closed-Loop System from Hypothesis to Verification. arXiv preprint arXiv:2505.16938

    2025 arXiv

  24. [32]

    Team, Q. 2025. QwQ-32B: Embracing the Power of Reinforcement Learning

    2025

  25. [33]

    Wen, M.; Kuba, J.; Lin, R.; Zhang, W.; Wen, Y.; Wang, J.; and Yang, Y. 2022. Multi-agent reinforcement learning is a sequence modeling problem. Advances in Neural Information Processing Systems, 35: 16509--16521

    2022

  26. [34]

    Wo \'z niak, S.; Koptyra, B.; Janz, A.; Kazienko, P.; and Koco \'n , J. 2024. Personalized large language models. arXiv preprint arXiv:2402.09269

    2024 arXiv

  27. [35]

    Y.; Woodruff, T

    Yang, Y.; Tian, T. Y.; Woodruff, T. K.; Jones, B. F.; and Uzzi, B. 2022. Gender-diverse teams produce more novel and higher-impact scientific ideas. Proceedings of the National Academy of Sciences, 119(36): e2200841119

    2022

  28. [36]

    Zeng, A.; Fan, Y.; Di, Z.; Wang, Y.; and Havlin, S. 2021. Fresh teams are associated with original and multidisciplinary research. Nature human behaviour, 5(10): 1314--1322

    2021

  29. [37]

    Zhou, Y.; Shen, T.; Geng, X.; Tao, C.; Shen, J.; Long, G.; Xu, C.; and Jiang, D. 2024. Fine-grained distillation for long document retrieval. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, 19732--19740

    2024

Pith tools

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