pith. sign in

arxiv: 2605.22878 · v1 · pith:3ATE34ICnew · submitted 2026-05-20 · 💻 cs.AI · cs.CL· cs.IR· cs.LG

SciAtlas: A Large-Scale Knowledge Graph for Automated Scientific Research

Shuofei Qiao , Yunxiang Wei , Jiazheng Fan , Bin Wu , Busheng Zhang , Mengru Wang , Yuqi Zhu , Ningyu Zhang
show 3 more authors
Keyan Ding Qiang Zhang Huajun Chen
This is my paper

Pith reviewed 2026-05-25 05:44 UTC · model grok-4.3

classification 💻 cs.AI cs.CLcs.IRcs.LG
keywords knowledge graphscientific literatureAI agentsneuro-symbolic retrievalinterdisciplinary researchautomated discoveryacademic retrievaltopological reasoning
0
0 comments X

The pith

SciAtlas builds a knowledge graph from 43 million papers to give AI agents a topological map of science for deterministic cross-discipline discovery.

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

Current academic search tools depend on keyword matching or vector similarity and therefore cannot trace logical connections that cross fields, which produces hallucinations when AI agents attempt deep research. The paper presents SciAtlas as a single heterogeneous graph assembled from 43 million papers across 26 disciplines, containing 157 million entities and 3 billion triplets. This graph is offered as a structured substrate that removes disciplinary boundaries and supplies AI agents with a global view of scientific evolution. A neuro-symbolic retrieval method that combines tri-path recall with graph reranking is shown to convert ordinary semantic matches into reliable association paths. The authors illustrate the graph's use in literature review, trend synthesis, idea placement, and trajectory mapping to argue that it can close the loop of automated research while lowering inference cost.

Core claim

SciAtlas is a large-scale, multi-disciplinary academic knowledge graph built from over 43 million papers that yields 157 million entities and 3 billion triplets and is structured as a panoramic scientific evolution network. It supplies a topological cognitive substrate that dismantles disciplinary barriers and equips AI agents with a global perspective. The accompanying neuro-symbolic retrieval algorithm, which performs tri-path collaborative recall followed by graph reranking, moves retrieval from simple semantic matching to deterministic association discovery and thereby supports the full cycle of automated scientific research at reduced reasoning cost.

What carries the argument

The neuro-symbolic retrieval algorithm that performs tri-path collaborative recall and graph reranking on the SciAtlas knowledge graph to convert semantic matches into deterministic cross-entity associations.

If this is right

  • Literature review can incorporate topological paths that link ideas across disciplines rather than isolated keyword hits.
  • Automated synthesis of research trends becomes possible by traversing the graph's evolution network instead of aggregating isolated papers.
  • Idea positioning can be performed by locating a new concept relative to existing association chains in the global map.
  • Academic trajectory exploration can follow deterministic sequences of entities and relations rather than statistical similarity alone.
  • Reasoning costs for agentic research frameworks decrease because the graph supplies explicit associations that replace open-ended inference steps.

Where Pith is reading between the lines

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

  • If the graph maintains accuracy at scale, it could serve as a shared substrate that multiple independent AI research agents query without each rebuilding its own knowledge base.
  • Periodic re-extraction from newly published papers would be required to keep association paths current; without updates the deterministic advantage would erode over time.
  • The same structure might be used to measure the density of cross-disciplinary links in any given subfield by counting shortest paths between entities from different disciplines.
  • Natural-language interfaces layered on the retrieval algorithm could let non-expert users pose complex multi-hop questions that resolve to verifiable graph paths.

Load-bearing premise

Automatic extraction of 157 million entities and 3 billion triplets from 43 million papers yields an accurate and unbiased representation of scientific knowledge that AI agents can apply directly without introducing logical errors.

What would settle it

A head-to-head test on a fixed set of interdisciplinary research queries that measures whether AI agents using SciAtlas retrieval produce fewer logical hallucinations and higher factual accuracy than the same agents using only vector semantic search.

read the original abstract

The exponential growth of global academic output has confronted researchers and AI agents with an unprecedented ``information explosion,'' where fragmented and unstructured knowledge organization impedes deep interdisciplinary integration. Current academic retrieval tools predominantly rely on superficial keyword matching or vector-space semantic retrieval, which lack the topological reasoning capabilities required to navigate complex logical connections. Agentic deep-research-based frameworks are often prone to logical hallucinations and consuming high inference costs. To bridge this gap, in this report, we introduce SciAtlas, a large-scale, multi-disciplinary, heterogeneous academic resource knowledge graph designed as a panoramic scientific evolution network. By integrating over 43M papers from 26 disciplines, and a total of 157M entities and 3B triplets, SciAtlas provides a structured topological cognitive substrate that dismantles disciplinary barriers and furnishes AI agents with a global perspective. Furthermore, we develop a neuro-symbolic retrieval algorithm featuring tri-path collaborative recall and graph reranking, achieving a seamless transition from simple semantic matching to deterministic association discovery. We also present key application directions of SciAtlas, including literature review, automated research trend synthesis, idea positioning, and academic trajectory exploration, to demonstrate that SciAtlas can serve as an effective ``cognitive map'' to empower the full loop of automated scientific research while significantly reducing reasoning costs. We have released the interfaces for KG retrieval and various downstream tasks in our GitHub repo.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper claims to introduce SciAtlas, a large-scale heterogeneous academic knowledge graph integrating 43 million papers from 26 disciplines, resulting in 157 million entities and 3 billion triplets. It presents a neuro-symbolic retrieval algorithm with tri-path collaborative recall and graph reranking that transitions from semantic matching to deterministic association discovery. The KG is positioned as a 'structured topological cognitive substrate' that enables AI agents to perform automated scientific research tasks like literature review and trend synthesis while reducing logical hallucinations and inference costs. Interfaces for retrieval and downstream tasks are released via GitHub.

Significance. If the accuracy of the automatic KG construction and the effectiveness of the retrieval algorithm are demonstrated, SciAtlas could provide a valuable panoramic view of scientific knowledge, facilitating interdisciplinary research and more reliable agent-based scientific discovery. This would address key limitations in current academic search tools and agentic frameworks.

major comments (2)
  1. [Abstract] Abstract: The abstract asserts the scale, the retrieval performance, and the reduction in reasoning costs but supplies no quantitative metrics, ablation studies, error analysis, or validation against baselines; the central claims therefore rest on assertion rather than demonstrated evidence.
  2. [KG extraction pipeline] KG extraction pipeline: No precision, recall, or error analysis is reported for the automatic extraction of 157M entities and 3B triplets, which is load-bearing for the claim that the KG serves as an accurate, unbiased substrate usable directly by AI agents without introducing logical hallucinations.
minor comments (1)
  1. [Abstract] Abstract: The phrase 'seamless transition from simple semantic matching to deterministic association discovery' is used without specifying the mechanism or providing supporting details.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the thoughtful and constructive report. The comments highlight important areas where the presentation of evidence can be strengthened. We respond to each major comment below and commit to revisions that directly address the concerns raised.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The abstract asserts the scale, the retrieval performance, and the reduction in reasoning costs but supplies no quantitative metrics, ablation studies, error analysis, or validation against baselines; the central claims therefore rest on assertion rather than demonstrated evidence.

    Authors: We agree that the abstract would be strengthened by including concrete quantitative support for the central claims. The body of the manuscript reports retrieval metrics, ablation results, and baseline comparisons in the experimental sections; however, these are not summarized in the abstract. In the revised version we will add a concise statement of key performance figures (e.g., recall@K improvements and inference-cost reductions) together with pointers to the relevant evaluation sections. revision: yes

  2. Referee: [KG extraction pipeline] KG extraction pipeline: No precision, recall, or error analysis is reported for the automatic extraction of 157M entities and 3B triplets, which is load-bearing for the claim that the KG serves as an accurate, unbiased substrate usable directly by AI agents without introducing logical hallucinations.

    Authors: The referee correctly identifies that a dedicated error analysis of the full extraction pipeline is absent. Because exhaustive manual validation at this scale is impractical, we relied on established extraction components whose accuracies are documented in the cited literature and performed limited spot-checks on sampled subgraphs. We will add a new subsection that reports the sampling-based validation protocol, the resulting precision/recall estimates, and an explicit discussion of residual risks of hallucination or bias, thereby making the supporting evidence transparent. revision: yes

Circularity Check

0 steps flagged

No circularity: construction paper with no load-bearing derivations or self-citation chains

full rationale

The paper describes the assembly of SciAtlas from 43M papers yielding 157M entities and 3B triplets, followed by a neuro-symbolic retrieval algorithm and example applications. No equations, fitted parameters, or predictions are presented that reduce by construction to the paper's own inputs. No self-citations are invoked as load-bearing uniqueness theorems or ansatzes. The central claims rest on the existence and utility of the constructed resource rather than any closed derivation loop, making the work self-contained against external benchmarks of KG construction and retrieval.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Construction details, entity extraction rules, relation schemas, and any validation procedures are absent from the abstract; therefore the ledger records only the high-level domain assumptions required to interpret the stated scale and purpose.

axioms (1)
  • domain assumption Academic papers can be automatically parsed into a heterogeneous knowledge graph of entities and triplets that faithfully represents scientific knowledge across disciplines.
    This premise is required for the claim that the resulting 157 M entities and 3 B triplets constitute a usable cognitive substrate.

pith-pipeline@v0.9.0 · 5810 in / 1390 out tokens · 24872 ms · 2026-05-25T05:44:11.796865+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

62 extracted references · 62 canonical work pages · 9 internal anchors

  1. [1]

    Reasoning with Language Model Prompting:

    Shuofei Qiao and Yixin Ou and Ningyu Zhang and Xiang Chen and Yunzhi Yao and Shumin Deng and Chuanqi Tan and Fei Huang and Huajun Chen , editor =. Reasoning with Language Model Prompting:. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers),. 2023 , url =. doi:10.18653/V1/2023.ACL-LONG.294 , timestamp =

    work page doi:10.18653/v1/2023.acl-long.294 2023

  2. [2]

    Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models

    Qiguang Chen and Libo Qin and Jinhao Liu and Dengyun Peng and Jiannan Guan and Peng Wang and Mengkang Hu and Yuhang Zhou and Te Gao and Wanxiang Che , title =. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2503.09567 , eprinttype =. 2503.09567 , timestamp =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2503.09567 2025

  3. [3]

    2025 , eprint=

    DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models , author=. 2025 , eprint=

    work page 2025

  4. [4]

    Introducing GPT-5 , note=

    OpenAI , year=. Introducing GPT-5 , note=

    work page

  5. [5]

    Qwen3 Technical Report

    An Yang and Anfeng Li and Baosong Yang and Beichen Zhang and Binyuan Hui and Bo Zheng and Bowen Yu and Chang Gao and Chengen Huang and Chenxu Lv and Chujie Zheng and Dayiheng Liu and Fan Zhou and Fei Huang and Feng Hu and Hao Ge and Haoran Wei and Huan Lin and Jialong Tang and Jian Yang and Jianhong Tu and Jianwei Zhang and Jian Yang and et al. , title =....

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2505.09388 2025

  6. [6]

    AI4Research:

    Qiguang Chen and Ming. AI4Research:. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2507.01903 , eprinttype =. 2507.01903 , timestamp =

    work page doi:10.48550/arxiv.2507.01903 2025

  7. [7]

    Agent Laboratory: Using LLM Agents as Research Assistants

    Samuel Schmidgall and Yusheng Su and Ze Wang and Ximeng Sun and Jialian Wu and Xiaodong Yu and Jiang Liu and Zicheng Liu and Emad Barsoum , title =. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2501.04227 , eprinttype =. 2501.04227 , timestamp =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2501.04227 2025

  8. [8]

    The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery

    Chris Lu and Cong Lu and Robert Tjarko Lange and Jakob N. Foerster and Jeff Clune and David Ha , title =. CoRR , volume =. 2024 , url =. doi:10.48550/ARXIV.2408.06292 , eprinttype =. 2408.06292 , timestamp =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2408.06292 2024

  9. [9]

    CoRR , volume =

    Qiujie Xie and Yixuan Weng and Minjun Zhu and Fuchen Shen and Shulin Huang and Zhen Lin and Jiahui Zhou and Zilan Mao and Zijie Yang and Linyi Yang and Jian Wu and Yue Zhang , title =. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2507.23276 , eprinttype =. 2507.23276 , timestamp =

    work page doi:10.48550/arxiv.2507.23276 2025

  10. [10]

    Weld and Doug Downey and Wen

    Akari Asai and Jacqueline He and Rulin Shao and Weijia Shi and Amanpreet Singh and Joseph Chee Chang and Kyle Lo and Luca Soldaini and Sergey Feldman and Mike D'Arcy and David Wadden and Matt Latzke and Minyang Tian and Pan Ji and Shengyan Liu and Hao Tong and Bohao Wu and Yanyu Xiong and Luke Zettlemoyer and Graham Neubig and Daniel S. Weld and Doug Down...

    work page doi:10.48550/arxiv.2411.14199 2024

  11. [11]

    CoRR , volume =

    Xun Liang and Jiawei Yang and Yezhaohui Wang and Chen Tang and Zifan Zheng and Shichao Song and Zehao Lin and Yebin Yang and Simin Niu and Hanyu Wang and Bo Tang and Feiyu Xiong and Keming Mao and Zhiyu Li , title =. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2502.14776 , eprinttype =. 2502.14776 , timestamp =

    work page doi:10.48550/arxiv.2502.14776 2025

  12. [12]

    AutoSurvey: Large Language Models Can Automatically Write Surveys , booktitle =

    Yidong Wang and Qi Guo and Wenjin Yao and Hongbo Zhang and Xin Zhang and Zhen Wu and Meishan Zhang and Xinyu Dai and Min Zhang and Qingsong Wen and Wei Ye and Shikun Zhang and Yue Zhang , editor =. AutoSurvey: Large Language Models Can Automatically Write Surveys , booktitle =. 2024 , url =

    work page 2024

  13. [13]

    Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers),

    Xiangchao Yan and Shiyang Feng and Jiakang Yuan and Renqiu Xia and Bin Wang and Lei Bai and Bo Zhang , editor =. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers),. 2025 , url =

    work page 2025

  14. [14]

    Laradji and Krishnamurthy Dj Dvijotham and Jason Stanley and Laurent Charlin and Christopher Pal , title =

    Shubham Agarwal and Gaurav Sahu and Abhay Puri and Issam H. Laradji and Krishnamurthy Dj Dvijotham and Jason Stanley and Laurent Charlin and Christopher Pal , title =. Trans. Mach. Learn. Res. , volume =. 2025 , url =

    work page 2025

  15. [15]

    CoRR , volume =

    Keyu Zhao and Weiquan Lin and Qirui Zheng and Fengli Xu and Yong Li , title =. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2511.02238 , eprinttype =. 2511.02238 , timestamp =

    work page doi:10.48550/arxiv.2511.02238 2025

  16. [16]

    ResearchAgent: Iterative Research Idea Generation over Scientific Literature with Large Language Models , booktitle =

    Jinheon Baek and Sujay Kumar Jauhar and Silviu Cucerzan and Sung Ju Hwang , editor =. ResearchAgent: Iterative Research Idea Generation over Scientific Literature with Large Language Models , booktitle =. 2025 , url =. doi:10.18653/V1/2025.NAACL-LONG.342 , timestamp =

    work page doi:10.18653/v1/2025.naacl-long.342 2025

  17. [17]

    CoRR , volume =

    Long Li and Weiwen Xu and Jiayan Guo and Ruochen Zhao and Xingxuan Li and Yuqian Yuan and Boqiang Zhang and Yuming Jiang and Yifei Xin and Ronghao Dang and Deli Zhao and Yu Rong and Tian Feng and Lidong Bing , title =. CoRR , volume =. 2024 , url =. doi:10.48550/ARXIV.2410.13185 , eprinttype =. 2410.13185 , timestamp =

    work page doi:10.48550/arxiv.2410.13185 2024

  18. [18]

    Many Heads Are Better Than One: Improved Scientific Idea Generation by

    Haoyang Su and Renqi Chen and Shixiang Tang and Zhenfei Yin and Xinzhe Zheng and Jinzhe Li and Biqing Qi and Qi Wu and Hui Li and Wanli Ouyang and Philip Torr and Bowen Zhou and Nanqing Dong , editor =. Many Heads Are Better Than One: Improved Scientific Idea Generation by. Proceedings of the 63rd Annual Meeting of the Association for Computational Lingui...

    work page 2025

  19. [19]

    CoRR , volume =

    Wenxiao Wang and Lihui Gu and Liye Zhang and Yunxiang Luo and Yi Dai and Chen Shen and Liang Xie and Binbin Lin and Xiaofei He and Jieping Ye , title =. CoRR , volume =. 2024 , url =. doi:10.48550/ARXIV.2410.23166 , eprinttype =. 2410.23166 , timestamp =

    work page doi:10.48550/arxiv.2410.23166 2024

  20. [20]

    AlphaEvolve: A coding agent for scientific and algorithmic discovery

    Alexander Novikov and Ng. AlphaEvolve:. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2506.13131 , eprinttype =. 2506.13131 , timestamp =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2506.13131 2025

  21. [21]

    CoRR , volume =

    Yixin Ou and Yujie Luo and Jingsheng Zheng and Lanning Wei and Shuofei Qiao and Jintian Zhang and Da Zheng and Huajun Chen and Ningyu Zhang , title =. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2506.10974 , eprinttype =. 2506.10974 , timestamp =

    work page doi:10.48550/arxiv.2506.10974 2025

  22. [22]

    arXiv preprint arXiv:2506.16499 (2025)

    Zexi Liu and Yuzhu Cai and Xinyu Zhu and Yujie Zheng and Runkun Chen and Ying Wen and Yanfeng Wang and Weinan E and Siheng Chen , title =. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2506.16499 , eprinttype =. 2506.16499 , timestamp =

    work page doi:10.48550/arxiv.2506.16499 2025

  23. [23]

    2025 , eprint=

    AlphaResearch: Accelerating New Algorithm Discovery with Language Models , author=. 2025 , eprint=

    work page 2025

  24. [24]

    AIDE: AI-Driven Exploration in the Space of Code

    Zhengyao Jiang and Dominik Schmidt and Dhruv Srikanth and Dixing Xu and Ian Kaplan and Deniss Jacenko and Yuxiang Wu , title =. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2502.13138 , eprinttype =. 2502.13138 , timestamp =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2502.13138 2025

  25. [25]

    2025 , eprint=

    XtraGPT: Context-Aware and Controllable Academic Paper Revision , author=. 2025 , eprint=

    work page 2025

  26. [26]

    CoRR , volume =

    Haomin Wen and Zhenjie Wei and Yan Lin and Jiyuan Wang and Yuxuan Liang and Huaiyu Wan , title =. CoRR , volume =. 2024 , url =. doi:10.48550/ARXIV.2403.09733 , eprinttype =. 2403.09733 , timestamp =

    work page doi:10.48550/arxiv.2403.09733 2024

  27. [27]

    The Thirteenth International Conference on Learning Representations,

    Yixuan Weng and Minjun Zhu and Guangsheng Bao and Hongbo Zhang and Jindong Wang and Yue Zhang and Linyi Yang , title =. The Thirteenth International Conference on Learning Representations,. 2025 , url =

    work page 2025

  28. [28]

    DeepReview: Improving LLM-based Paper Review with Human-like Deep Thinking Process , booktitle =

    Minjun Zhu and Yixuan Weng and Linyi Yang and Yue Zhang , editor =. DeepReview: Improving LLM-based Paper Review with Human-like Deep Thinking Process , booktitle =. 2025 , url =

    work page 2025

  29. [29]

    Reviewer2: Optimizing Review Generation Through Prompt Generation , journal =

    Zhaolin Gao and Kiant. Reviewer2: Optimizing Review Generation Through Prompt Generation , journal =. 2024 , url =. doi:10.48550/ARXIV.2402.10886 , eprinttype =. 2402.10886 , timestamp =

    work page doi:10.48550/arxiv.2402.10886 2024

  30. [30]

    AgentReview: Exploring Peer Review Dynamics with

    Yiqiao Jin and Qinlin Zhao and Yiyang Wang and Hao Chen and Kaijie Zhu and Yijia Xiao and Jindong Wang , editor =. AgentReview: Exploring Peer Review Dynamics with. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing,. 2024 , url =. doi:10.18653/V1/2024.EMNLP-MAIN.70 , timestamp =

    work page doi:10.18653/v1/2024.emnlp-main.70 2024

  31. [31]

    CoRR , volume =

    Jiabin Tang and Lianghao Xia and Zhonghang Li and Chao Huang , title =. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2505.18705 , eprinttype =. 2505.18705 , timestamp =

    work page doi:10.48550/arxiv.2505.18705 2025

  32. [32]

    2025 , eprint=

    OmniScientist: Toward a Co-evolving Ecosystem of Human and AI Scientists , author=. 2025 , eprint=

    work page 2025

  33. [33]

    CoRR , volume =

    Yixuan Weng and Minjun Zhu and Qiujie Xie and Qiyao Sun and Zhen Lin and Sifan Liu and Yue Zhang , title =. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2509.26603 , eprinttype =. 2509.26603 , timestamp =

    work page doi:10.48550/arxiv.2509.26603 2025

  34. [34]

    The AI Scientist-v2: Workshop-Level Automated Scientific Discovery via Agentic Tree Search

    Yutaro Yamada and Robert Tjarko Lange and Cong Lu and Shengran Hu and Chris Lu and Jakob N. Foerster and Jeff Clune and David Ha , title =. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2504.08066 , eprinttype =. 2504.08066 , timestamp =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2504.08066 2025

  35. [35]

    CoRR , volume =

    Yao Wang and Mingxuan Cui and Arthur Jiang , title =. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2503.01508 , eprinttype =. 2503.01508 , timestamp =

    work page doi:10.48550/arxiv.2503.01508 2025

  36. [36]

    The Thirteenth International Conference on Learning Representations,

    Chenglei Si and Diyi Yang and Tatsunori Hashimoto , title =. The Thirteenth International Conference on Learning Representations,. 2025 , url =

    work page 2025

  37. [37]

    The Thirteenth International Conference on Learning Representations,

    Tao Feng and Yihang Sun and Jiaxuan You , title =. The Thirteenth International Conference on Learning Representations,. 2025 , url =

    work page 2025

  38. [38]

    CoRR , volume =

    Bo Zhang and Shiyang Feng and Xiangchao Yan and Jiakang Yuan and Zhiyin Yu and Xiaohan He and Songtao Huang and Shaowei Hou and Zheng Nie and Zhilong Wang and Jinyao Liu and Runmin Ma and Tianshuo Peng and Peng Ye and Dongzhan Zhou and Shufei Zhang and Xiaosong Wang and Yilan Zhang and Meng Li and Zhongying Tu and Xiangyu Yue and Wangli Ouyang and Bowen Z...

    work page doi:10.48550/arxiv.2505.16938 2025

  39. [39]

    ScholarEval: Research Idea Evaluation Grounded in Literature , journal =

    Hanane Nour Moussa and Patrick Queiroz Da Silva and Daniel Adu. ScholarEval: Research Idea Evaluation Grounded in Literature , journal =. 2025 , url =. doi:10.48550/ARXIV.2510.16234 , eprinttype =. 2510.16234 , timestamp =

    work page doi:10.48550/arxiv.2510.16234 2025

  40. [40]

    C-Pack: Packed Resources For General Chinese Embeddings , booktitle =

    Shitao Xiao and Zheng Liu and Peitian Zhang and Niklas Muennighoff and Defu Lian and Jian. C-Pack: Packed Resources For General Chinese Embeddings , booktitle =. 2024 , url =. doi:10.1145/3626772.3657878 , timestamp =

    work page doi:10.1145/3626772.3657878 2024

  41. [41]

    Introducing OpenAI o3 and o4-mini , note=

    OpenAI , year=. Introducing OpenAI o3 and o4-mini , note=

    work page

  42. [42]

    2026 , eprint=

    OpenNovelty: An LLM-powered Agentic System for Verifiable Scholarly Novelty Assessment , author=. 2026 , eprint=

    work page 2026

  43. [43]

    Weld and Tom Hope , title =

    Simra Shahid and Marissa Radensky and Raymond Fok and Pao Siangliulue and Daniel S. Weld and Tom Hope , title =. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2506.22026 , eprinttype =. 2506.22026 , timestamp =

    work page doi:10.48550/arxiv.2506.22026 2025

  44. [44]

    CoRR , volume =

    Jingyi Chai and Shuo Tang and Rui Ye and Yuwen Du and Xinyu Zhu and Mengcheng Zhou and Yanfeng Wang and Weinan E and Yuzhi Zhang and Linfeng Zhang and Siheng Chen , title =. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2507.05241 , eprinttype =. 2507.05241 , timestamp =

    work page doi:10.48550/arxiv.2507.05241 2025

  45. [45]

    Mamba: Linear-Time Sequence Modeling with Selective State Spaces

    Albert Gu and Tri Dao , title =. CoRR , volume =. 2023 , url =. doi:10.48550/ARXIV.2312.00752 , eprinttype =. 2312.00752 , timestamp =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2312.00752 2023

  46. [46]

    Efficiently Modeling Long Sequences with Structured State Spaces , booktitle =

    Albert Gu and Karan Goel and Christopher R. Efficiently Modeling Long Sequences with Structured State Spaces , booktitle =. 2022 , url =

    work page 2022

  47. [47]

    The Twelfth International Conference on Learning Representations,

    Tri Dao , title =. The Twelfth International Conference on Learning Representations,. 2024 , url =

    work page 2024

  48. [48]

    Fu and Tri Dao and Khaled Kamal Saab and Armin W

    Daniel Y. Fu and Tri Dao and Khaled Kamal Saab and Armin W. Thomas and Atri Rudra and Christopher R. Hungry Hungry Hippos: Towards Language Modeling with State Space Models , booktitle =. 2023 , url =

    work page 2023

  49. [49]

    Nature Reviews Psychology , volume=

    Information aggregation and collective intelligence beyond the wisdom of crowds , author=. Nature Reviews Psychology , volume=. 2022 , publisher=

    work page 2022

  50. [50]

    Nature Human Behaviour , volume=

    Aggregated knowledge from a small number of debates outperforms the wisdom of large crowds , author=. Nature Human Behaviour , volume=. 2018 , publisher=

    work page 2018

  51. [51]

    Impact of urbanization on water shortage in face of climatic aberrations , pages=

    Multi criteria decision making , author=. Impact of urbanization on water shortage in face of climatic aberrations , pages=. 2015 , publisher=

    work page 2015

  52. [52]

    Journal of knowledge management , volume=

    Innovation as a knowledge-based outcome , author=. Journal of knowledge management , volume=. 2011 , publisher=

    work page 2011

  53. [53]

    Journal of knowledge management , volume=

    The role of knowledge management in innovation , author=. Journal of knowledge management , volume=. 2007 , publisher=

    work page 2007

  54. [54]

    CoRR , volume =

    Shuofei Qiao and Yanqiu Zhao and Zhisong Qiu and Xiaobin Wang and Jintian Zhang and Zhao Bin and Ningyu Zhang and Yong Jiang and Pengjun Xie and Fei Huang and Huajun Chen , title =. CoRR , volume =. 2025 , url =. doi:10.48550/ARXIV.2509.25084 , eprinttype =. 2509.25084 , timestamp =

    work page doi:10.48550/arxiv.2509.25084 2025

  55. [55]

    2026 , eprint=

    DSGym: A Holistic Framework for Evaluating and Training Data Science Agents , author=. 2026 , eprint=

    work page 2026

  56. [56]

    WisPaper: Your AI Scholar Search Engine

    Li Ju and Jun Zhao and Mingxu Chai and Ziyu Shen and Xiangyang Wang and Yage Geng and Chunchun Ma and Hao Peng and Guangbin Li and Tao Li and Chengyong Liao and Fu Wang and Xiaolong Wang and Junshen Chen and Rui Gong and Shijia Liang and Feiyan Li and Ming Zhang and Kexin Tan and Jujie Ye and Zhiheng Xi and Shihan Dou and Tao Gui and Yuankai Ying and Yang...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2512.06879 2025

  57. [57]

    Rahmani and Yanshan Wang and Qiang Zhang and Keyan Ding and Jeff Z

    Shuofei Qiao and Yunxiang Wei and Xuehai Wang and Bin Wu and Boyang Xue and Ningyu Zhang and Hossein A. Rahmani and Yanshan Wang and Qiang Zhang and Keyan Ding and Jeff Z. Pan and Huajun Chen and Emine Yilmaz , title =. CoRR , volume =. 2026 , url =. doi:10.48550/ARXIV.2602.14367 , eprinttype =. 2602.14367 , timestamp =

    work page doi:10.48550/arxiv.2602.14367 2026

  58. [58]

    CoRR , volume =

    Hongjin Qian and Ziyi Xia and Ze Liu and Jianlyu Chen and Kun Luo and Minghao Qin and Chaofan Li and Lei Xiong and Junwei Lan and Sen Wang and Zhengyang Liang and Yingxia Shao and Defu Lian and Zheng Liu , title =. CoRR , volume =. 2026 , url =. doi:10.48550/ARXIV.2603.00084 , eprinttype =. 2603.00084 , timestamp =

    work page doi:10.48550/arxiv.2603.00084 2026

  59. [59]

    ArxivQA: Training Retrieval Agents for arXiv Search , note=

    Rehaan Ahmad and Daniel Kim , year=. ArxivQA: Training Retrieval Agents for arXiv Search , note=

    work page

  60. [60]

    Fast Random Walk with Restart and Its Applications , booktitle =

    Hanghang Tong and Christos Faloutsos and Jia. Fast Random Walk with Restart and Its Applications , booktitle =. 2006 , url =. doi:10.1109/ICDM.2006.70 , timestamp =

    work page doi:10.1109/icdm.2006.70 2006

  61. [61]

    Bridging Data and Discovery: A Survey on Knowledge Graphs in AI for Science , url=

    Ding, Keyan and Zhu, Zhihui and Tang, Yuqi and Feng, Kehua and Zhuang, Xiang and Wang, Hongwei and Yang, Yi and Du, Huifang and Ni, Zhangkai and Wang, Shiqi and Fan, Xiaohui and Xing, Huabin and Bai, Lei and Liu, Qi and Wang, Haofen and Zhang, Qiang and Chen, Huajun , year=. Bridging Data and Discovery: A Survey on Knowledge Graphs in AI for Science , url...

    work page doi:10.36227/techrxiv.176369442.22009541/v1

  62. [62]

    2026 , eprint=

    Evaluating LLMs' Divergent Thinking Capabilities for Scientific Idea Generation with Minimal Context , author=. 2026 , eprint=

    work page 2026