Pith. sign in

REVIEW 4 major objections 5 minor 2 cited by

HiPerRAG: High-Performance Retrieval Augmented Generation for Scientific Insights

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

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

Pith's one-line read HiPerRAG shows that a retrieval pipeline over 3.6 million scientific papers can lift a general-purpose LLM to 90% on SciQ and 76% on PubMedQA, beating domain-specific and commercial baselines.

desk verdict Solid engineering on scalable scientific RAG, but the abstract's headline numbers come from different configurations and the novel retriever sometimes hurts QA accuracy. read the letter →

arxiv 2505.04846 v1 pith:CNY2Q3ZH submitted 2025-05-07 cs.IR cs.CEcs.CLcs.DCcs.LG

Ozan Gokdemir , Carlo Siebenschuh , Alexander Brace , Azton Wells , Brian Hsu , Kyle Hippe , Priyanka V. Setty , Aswathy Ajith
show 16 more authors
J. Gregory Pauloski Varuni Sastry Sam Foreman Huihuo Zheng Heng Ma Bharat Kale Nicholas Chia Thomas Gibbs Michael E. Papka Thomas Brettin Francis J. Alexander Anima Anandkumar Ian Foster Rick Stevens Venkatram Vishwanath Arvind Ramanathan
This is my paper · ORCID
classification cs.IRcs.CEcs.CLcs.DCcs.LG
keywords retrieval-augmentedgenerationscientificquestionansweringencoderfine-tuningcontrastivelearninglateinteractiondocumentparsinghigh-performancecomputingbiomedicalbenchmarks
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

HiPerRAG claims that retrieval-augmented generation can be scaled to millions of scientific papers without sacrificing answer quality, by pairing a fast two-stage PDF parser (Oreo) with a query-aware encoder (ColTrast) fine-tuned through a combined contrastive and late-interaction loss. The system indexes more than 3.6 million articles and reports 90% accuracy on SciQ and 76% on PubMedQA, outperforming both a domain-specific biomedical model and a commercial general-purpose LLM on those benchmarks. The paper also introduces two biomedical question-answering benchmarks built from experimentally validated peptide-protein data, plus a synthetic retrieval-evaluation set, so that scientific RAG performance can be measured beyond general-knowledge tests. If the claims hold, the practical consequence is that a general-purpose LLM equipped with retrieval from a large scientific corpus can match or beat models trained specifically for science, without fine-tuning the generator.

What carries the argument

The load-bearing mechanism is the ColTrast loss, a training objective that averages two terms: a contrastive loss that pulls the pooled embedding of a question close to its matching chunk and pushes it away from other chunks, and a late-interaction max-similarity loss that compares each query token against document tokens for a finer-grained match. Because the late-interaction term is computed only on the local GPU while the contrastive term gathers embeddings across all GPUs, the method keeps scalability without losing token-level alignment. Oreo is the second mechanism: it first detects layout regions with a convolutional network and then decodes only the text regions with a vision-transformer decoder, which is what gives the reported throughput gain.

What would settle it

Take a set of questions written by working researchers about the same corpus, without letting the system see the passages first; if retrieval precision or end-to-end answer accuracy falls to the level of the unfine-tuned base encoder, the ColTrast advantage is an artifact of training on questions generated from the target chunks.

Watch

Extended reading notes

Core claim

The central claim is that the two main bottlenecks to million-document scientific retrieval-augmented generation, parsing PDFs and aligning query and passage embeddings, can both be overcome at scale. Oreo splits parsing into layout detection followed by targeted text decoding, achieving roughly 4.5 times higher throughput than the prior parser it compares against while reaching comparable or better character accuracy. ColTrast fine-tunes a general embedding model by averaging a contrastive loss over pooled query-document embeddings with a late-interaction max-similarity loss computed locally, so that retrieval benefits from fine-grained token alignment while remaining efficient enough for a vector database. On the assembled corpus of 3.6 million papers, the best configuration answers 90% of SciQ questions and 76% of PubMedQA questions, and the paper attributes this to retrieval supplying the generator with the relevant passages rather than to generator fine-tuning.

Load-bearing premise

The load-bearing assumption is that the 455,894 questions generated by a large language model from the same semantic chunks the system must retrieve are representative of how real scientists ask questions; if that distribution is off, the measured retrieval and downstream question-answering gains will not transfer to actual use.

Editorial extensions

If this is right

  • The retrieval step, not generator fine-tuning, is what the paper credits for the accuracy gains: the same instruction-tuned generator without retrieval scores 78% on SciQ, and adding ColTrast retrieval over the full corpus brings it to 90%.
  • ColTrast's combined loss beats contrastive-only and late-interaction-only training on the paper's retrieval metrics, so the two losses together are presented as a reusable recipe for domain-specific retrieval.
  • The two new protein QA benchmarks and the synthetic retrieval set evaluate RAG on experimentally verified peptide-protein relationships, giving the community tests that go beyond general encyclopedic facts.
  • The workflow's linear scaling on GPU clusters means the parsing, chunking, and embedding stages can in principle be pointed at the roughly 200 million published papers, not just the 3.6 million indexed here.

Reading between the lines

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

  • Because ColTrast is trained on questions generated from the same chunks it must later retrieve, the reported retrieval gains could shrink on naturally occurring queries; a test with questions written by working scientists would separate genuine encoder quality from training-set overlap.
  • The 90% SciQ figure is a system-level result: the generator alone scores 78% on SciQ, so part of the gain is simply the value of retrieving a directly relevant passage, and how much ColTrast adds over a cheaper encoder on that benchmark remains an open question.
  • The recipe likely transfers to other fields: generate question-chunk pairs from a domain corpus, fine-tune a general encoder with the combined loss, and index that same corpus, provided the synthetic questions match how the field's researchers actually phrase queries.
  • The embeddings produced by ColTrast could plausibly be reused as a similarity graph over scientific concepts, supporting literature-mapping and hypothesis-generation tools, though the paper only sketches that direction.
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 / 5 minor

Summary. The paper presents HiPerRAG, a distributed retrieval-augmented generation workflow for scientific literature, combining an HPC orchestration layer with three technical contributions: Oreo, a layout-aware PDF parser; ColTrast, a query-aware encoder fine-tuning method that mixes contrastive and late-interaction losses; and two new QA benchmarks (ProteinInteractionQA, ProteinFunctionQA) plus a synthetic retrieval dataset (BioSynthQP). The authors claim that HiPerRAG achieves 90% accuracy on SciQ and 76% on PubMedQA, outperforming PubMedGPT and GPT-4, and that it scales to thousands of GPUs on Polaris, Sunspot, and Frontier. The evaluation also reports Oreo's throughput advantages and strong-scaling behavior of parsing, semantic chunking, and encoder fine-tuning.

Significance. If the headline results held, HiPerRAG would be a practically important step toward million-document scientific RAG, and the scaling measurements plus the new benchmarks would be useful community resources. The paper is honest about several engineering details, and the Oreo throughput results and scaling curves are concrete. However, the central accuracy claims are not supported by the paper's own Table 5: the 76% PubMedQA figure comes from a PubMedBERT baseline rather than ColTrast, and ColTrast-based retrieval often performs no better than, or worse than, no retrieval at all. The evaluation also relies heavily on synthetic question-generation pipelines that share a model family with the evaluated generators, so the external validity of the new benchmarks is uncertain.

major comments (4)
  1. [Abstract and Section 4.4, Table 5] The abstract's headline numbers are not reproducible with a single HiPerRAG configuration. The 76% PubMedQA result is obtained with Mixtral8x7B plus the off-the-shelf PubMedBERT encoder on the Protein Literature Corpus, not with either ColTrast encoder; the ColTrast-M-Q-S with SLC configuration scores 44% with Mistral and 75% with Mixtral. Similarly, the 90% SciQ result uses Mixtral8x7B plus ColTrast-B-S, while the same encoder with Mistral gives 82%, below the 88% no-retrieval Mistral baseline. The abstract should be rewritten to report per-configuration results and to define the specific system whose performance is being claimed.
  2. [Table 5 and Section 4.2] ColTrast retrieval degrades or only matches no-retrieval accuracy on most Mistral rows: SciQ 85 vs 88, LitQA 24 vs 40, ProteinInteractionQA 43 vs 44, and PubMedQA 44 vs 59. This contradicts the claim that ColTrast 'enhances retrieval accuracy' and that retrieval at scale improves end-to-end QA. The paper needs a direct retrieval-quality analysis on the actual QA evidence (e.g., hit rate of the answer-containing chunk in the retrieved top-K), a defined top-K, and a unified system configuration with error bars before the retrieval benefit can be accepted.
  3. [Sections 3.2.2 and 4.3.1/4.3.2] The evaluation is partially circular. ColTrast is fine-tuned on 455,894 Mistral-7B-generated questions derived from the same semantic chunks it must retrieve, and two of the five QA benchmarks are generated by Mistral-7B from UniProt records. Since Mistral/Mixtral are also the generators evaluated in Table 5, high scores on ProteinInteractionQA and ProteinFunctionQA may reflect distribution overlap between the training-time and evaluation-time question generators rather than retrieval quality. Please provide human evaluation, questions from independent sources, or an analysis showing that the generated question distribution matches real scientific information needs.
  4. [Section 4.4] The claimed superiority over GPT-4 and PubMedGPT is not supported by controlled experiments. Table 5 reports a single accuracy number per cell with no confidence intervals, no multiple seeds, and no statistical test; the PubMedQA advantage over GPT-4 is 0.8 percentage points. There is also no GPT-4 retrieval-augmented baseline run under the same protocol. Add uncertainty quantification and a same-protocol comparison before making comparative claims.
minor comments (5)
  1. [Section 4.2] The precision formula in the text is garbled; it appears as "P=N_rel/N_ret" with stray characters. Please typeset it cleanly and define N_rel and N_ret explicitly.
  2. [Sections 3.2 and 4.2] The encoder base model is referred to inconsistently as SFR-Mistral, SFR-Embedding-Mistral, and M in Table 4; please define the model name once and use it consistently.
  3. [Section 4.2] The 5% held-out evaluation set from the ColTrast training data is described only in passing; state its size, the document-level split procedure, and why it does not overlap with the training chunks.
  4. [Figure 5] The caption says 'Unless noted, a run was accomplished on Polaris,' but panel labels and the Frontier runs in panel C are ambiguous; clarify which system each curve corresponds to, and note that one Frontier GPU is one GCD.
  5. [References] References [33] and [34] are duplicates of the same PubMedBERT paper; also, spell 'PubMedQA' consistently.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the headline claims rest on external benchmarks and no derivation reduces to its inputs by construction.

full rationale

The paper's central claims (90% on SciQ, 76% on PubMedQA) are evaluations on established external benchmarks; neither benchmark appears in the ColTrast fine-tuning objective, so these numbers are not forced by construction. The ColTrast fine-tuning pipeline does generate 455,894 questions from its own semantic chunks (Section 3.2.2), and the Section 4.2 held-out evaluation draws from the same synthetic distribution, but the held-out pairs come from distinct documents and the model must generalize to unseen chunks, so this is a standard (if distributionally narrow) evaluation rather than a circular reduction. The two new QA benchmarks derive their correct answers from UniProt ground-truth annotations, not from the generator's free-form output, so their labels are not definitionally tied to the system being evaluated. The only author-overlapping citation (parsl_object_registry [16]) is a software tool reference and is not load-bearing for any theoretical claim. The internal-consistency concern that no single configuration reproduces both headline numbers is a reporting or cherry-picking issue, which is outside the scope of circularity analysis.

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

The reported system is empirical. It relies on the integrity of pre-trained encoders, on synthetic questions standing in for real user queries, on UniProt as ground truth, and on human-supervised GPT-4 labels in BioSynthQP. No free parameter is fitted to a target accuracy, but several hyperparameters that affect the results are chosen by hand or left unspecified.

free parameters (4)
  • ColTrast loss weight for late-interaction term = 0.5
    Section 3.2.1 sets the total loss as (L_LI + L_C)/2 with equal weighting and no ablation showing this choice is optimal.
  • Semantic chunking cosine similarity threshold
    Section 3.2.2 defines chunk boundaries by a predetermined threshold but does not report the value, even though chunk granularity controls what is retrieved.
  • Retrieval top-K context chunks
    The number of chunks supplied to the generator is not reported, and QA accuracy depends on this value.
  • QLoRA batch size per GPU = 24
    Section 3.2.1 chooses batch size 24 per GPU to fit one A100, and the effective number of contrastive negatives depends on this choice.
assumptions (4)
  • domain assumption Pre-trained encoders (SFR-Embedding-Mistral, BERT) provide valid semantic embeddings for scientific text before fine-tuning.
    Invoked throughout Section 3.2 as the starting point for ColTrast.
  • ad hoc to paper Mistral-7B-Instruct generated questions from semantic chunks are representative of real user queries for scientific QA.
    Section 3.2.2 uses 455,894 such questions as the sole fine-tuning supervision; no validation against human queries is reported.
  • domain assumption UniProt protein interaction and function records are complete and correct ground truth for the new QA benchmarks.
    Sections 4.3.1 and 4.3.2 use UniProt as ground truth; incomplete annotations could make randomly selected distractors accidentally correct.
  • domain assumption Human-supervised GPT-4-generated BioSynthQP relevance labels are accurate.
    Section 4.3.3 says human supervision guarantees label accuracy but gives no protocol, annotation counts, or inter-rater agreement.

how reviews work

0 comments
Cite this review

Pith. "Pith review of HiPerRAG: High-Performance Retrieval Augmented Generation for Scientific Insights." pith.science (2026). https://pith.science/paper/CNY2Q3ZH

@misc{pith2026250504846,
  author       = {Pith},
  title        = {Pith review of: HiPerRAG: High-Performance Retrieval Augmented Generation for Scientific Insights},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CNY2Q3ZH}},
  note         = {Machine review of arXiv:2505.04846}
}
read the original abstract

The volume of scientific literature is growing exponentially, leading to underutilized discoveries, duplicated efforts, and limited cross-disciplinary collaboration. Retrieval Augmented Generation (RAG) offers a way to assist scientists by improving the factuality of Large Language Models (LLMs) in processing this influx of information. However, scaling RAG to handle millions of articles introduces significant challenges, including the high computational costs associated with parsing documents and embedding scientific knowledge, as well as the algorithmic complexity of aligning these representations with the nuanced semantics of scientific content. To address these issues, we introduce HiPerRAG, a RAG workflow powered by high performance computing (HPC) to index and retrieve knowledge from more than 3.6 million scientific articles. At its core are Oreo, a high-throughput model for multimodal document parsing, and ColTrast, a query-aware encoder fine-tuning algorithm that enhances retrieval accuracy by using contrastive learning and late-interaction techniques. HiPerRAG delivers robust performance on existing scientific question answering benchmarks and two new benchmarks introduced in this work, achieving 90% accuracy on SciQ and 76% on PubMedQA-outperforming both domain-specific models like PubMedGPT and commercial LLMs such as GPT-4. Scaling to thousands of GPUs on the Polaris, Sunspot, and Frontier supercomputers, HiPerRAG delivers million document-scale RAG workflows for unifying scientific knowledge and fostering interdisciplinary innovation.

Figures

Figures reproduced from arXiv: 2505.04846 by the authors.

Figure 1
Figure 1. HiPerRAG Workflow. A graphical overview of our [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Parsing workflow for scientific PDFs. The neural [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Encoder/retrieval models workflow using the [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Sub-domain distribution of the antimicrobial pep [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: A) Strong scaling results for PDF parsing workflows on Polaris and Sunspot; B) Strong scaling results for semantic [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]

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. KinyaColBERT: A Lexically Grounded Retrieval Model for Low-Resource Retrieval-Augmented Generation

    cs.CL 2025-07 conditional novelty 5.0 of 10

    KinyaColBERT, a morphology-aware two-tier ColBERT retriever, reports large MRR gains over multilingual baselines and commercial APIs on a new Kinyarwanda agricultural retrieval benchmark.

  2. AI4Research: A Survey of Artificial Intelligence for Scientific Research

    cs.CL 2025-07 conditional novelty 4.0 of 10

    A survey that organizes AI-for-research work into five tasks, comprehension, survey, discovery, writing, and peer review, and compiles associated tools and benchmarks.

Reference graph

Works this paper leans on

88 extracted references · 28 canonical work pages · cited by 2 Pith papers

  1. [1]

    Enhancing Scholarly Data

    2016.Semantics, Analytics, Visualization. Enhancing Scholarly Data. Springer International Publishing. https://doi.org/10.1007/978-3-319-53637-8

    work page doi:10.1007/978-3-319-53637-8 2016

  2. [2]

    PubMedQA: A Dataset for Biomedical Research Question Answering

    2023. PubMedQA: A Dataset for Biomedical Research Question Answering. PubMedQA Project Website. https://pubmedqa.github.io/ Accessed: 2023-10-11

    2023

  3. [3]

    PDFMiner

    2024. PDFMiner. https://pypi.org/project/pdfminer/. Accessed: [12/7/2024]

    2024

  4. [4]

    PyMuPDF Documentation

    2024. PyMuPDF Documentation. https://pymupdf.readthedocs.io/en/latest/. Accessed: [12/7/2024]

    2024

  5. [5]

    PyPDF Documentation

    2024. PyPDF Documentation. https://pypdf.readthedocs.io/en/stable/. Accessed: [12/7/2024]

    2024

  6. [6]

    Zeeshan Ahmed and Thomas Dandekar. 2017. MSL: Facilitating automatic and physical analysis of published scientific literature in PDF format.F1000Research 4 (April 2017), 1453. https://doi.org/10.12688/f1000research.7329.2

    work page doi:10.12688/f1000research.7329.2 2017

  7. [7]

    Opeoluwa Akinseloyin, Xiaorui Jiang, and Vasile Palade. 2024. A question- answering framework for automated abstract screening using large language models.Journal of the American Medical Informatics Association31, 9 (July 2024), 1939–1952. https://doi.org/10.1093/jamia/ocae166

    doi:10.1093/jamia/ocae166 2024

  8. [8]

    Zhiyu An, Xianzhong Ding, Yen-Chun Fu, Cheng-Chung Chu, Yan Li, and Wan Du

Show all 88 references
  1. [9]

    Akari Asai, Zeqiu Wu, Yizhong Wang, Avirup Sil, and Hannaneh Hajishirzi. 2023. Self-RAG: Learning to retrieve, generate, and critique through self-reflection. arXiv preprint arXiv:2310.11511(2023)

    2023 arXiv

  2. [10]

    Sören Auer, Dante A. C. Barone, Cassiano Bartz, Eduardo G. Cortes, Mo- hamad Yaser Jaradeh, Oliver Karras, Manolis Koubarakis, Dmitry Mouromtsev, Dmitrii Pliukhin, Daniil Radyush, Ivan Shilin, Markus Stocker, and Eleni Tsalap- ati. 2023. The SciQA Scientific Question Answering...

    2023 doi

  3. [11]

    Yadu Babuji, Anna Woodard, Zhuozhao Li, Ben Clifford, Rohan Kumar, Lukasz Lacinski, Ryan Chard, Justin Wozniak, Ian Foster, Michael Wilde, Daniel Katz, and Kyle Chard. 2019. Parsl: Pervasive Parallel Programming in Python. InACM International Symposium on High-Performance Para...

    2019

  4. [12]

    Philip Bachman, R Devon Hjelm, and William Buchwalter. 2019. Learning Rep- resentations by Maximizing Mutual Information Across Views.arXiv preprint arXiv:1906.00910(2019)

    2019 arXiv

  5. [13]

    Sachin Banker, Promothesh Chatterjee, Himanshu Mishra, and Arul Mishra. 2024. Machine-assisted social psychology hypothesis generation.American Psychologist 79, 6 (Sept. 2024), 789–797. https://doi.org/10.1037/amp0001222

    2024 doi

  6. [14]

    Lukas Blecher, Guillem Cucurull, Thomas Scialom, and Robert Stojnic. 2023. Nougat: Neural optical understanding for academic documents.arXiv preprint arXiv:2308.13418(2023)

    2023 arXiv

  7. [15]

    Bolton, D

    E. Bolton, D. Hall, M. Yasunaga, T. Lee, C. Manning, and P. Liang. 2022. Stanford CRFM Introduces PubMedGPT 2.7B. Stanford HAI. https://hai.stanford.edu/ news/stanford-crfm-introduces-pubmedgpt-27b

    2022

  8. [16]

    Gregory Pauloski

    Alexander Brace and J. Gregory Pauloski. 2023. https://github.com/braceal/parsl_ object_registry. Accessed: 2024-10-09

    2023

  9. [17]

    Yapei Chang, Kyle Lo, Tanya Goyal, and Mohit Iyyer. 2023. BooookScore: A systematic exploration of book-length summarization in the era of LLMs. https: //doi.org/10.48550/ARXIV.2310.00785

    2023 doi arXiv

  10. [18]

    Kexin Chen, Junyou Li, Kunyi Wang, Yuyang Du, Jiahui Yu, Jiamin Lu, Lanqing Li, Jiezhong Qiu, Jianzhang Pan, Yi Huang, Qun Fang, Pheng Ann Heng, and Guangy- ong Chen. 2024. Chemist-X: Large Language Model-empowered Agent for Reac- tion Condition Recommendation in Chemical Synt...

    2024 arXiv

  11. [19]

    Wenhu Chen, Hexiang Hu, Xi Chen, Pat Verga, and William W. Cohen. 2022. MuRAG: Multimodal Retrieval-Augmented Generator for Open Question An- swering over Images and Text. arXiv:2210.02928 [cs.CL] https://arxiv.org/abs/ 2210.02928

    2022 arXiv

  12. [20]

    Johan SG Chu and James A Evans. 2021. Slowed canonical progress in large fields of science.Proceedings of the National Academy of Sciences118, 41 (2021), e2021636118

    2021

  13. [21]

    Yung-Sung Chuang, Yujia Xie, Hongyin Luo, Yoon Kim, James Glass, and Pengcheng He. 2023. Dola: Decoding by contrasting layers improves factuality in large language models.arXiv preprint arXiv:2309.03883(2023)

    2023 arXiv

  14. [22]

    2024.Science and Engineering Indicators 2024: The State of U.S

    Steven Deitz and Christina Freyman. 2024.Science and Engineering Indicators 2024: The State of U.S. Science and Engineering. Technical Report NSB-2024-3. National Science Foundation, Alexandria, VA. https://ncses.nsf.gov/pubs/nsb20243

    2024

  15. [23]

    Tim Dettmers, Artidoro Pagnoni, Ari Holtzman, and Luke Zettlemoyer. 2023. QLoRA: Efficient Finetuning of Quantized LLMs.arXiv preprint arXiv:2305.14314 (2023)

    2023 arXiv

  16. [24]

    Matthijs Douze, Alexandr Guzhva, Chengqi Deng, Jeff Johnson, Gergely Szilvasy, Pierre-Emmanuel Mazaré, Maria Lomeli, Lucas Hosseini, and Hervé Jégou. 2024. The Faiss library. arXiv:2401.08281 [cs.LG]

    2024 arXiv

  17. [25]

    Darren Edge, Ha Trinh, Newman Cheng, Joshua Bradley, Alex Chao, Apurva Mody, Steven Truitt, and Jonathan Larson. 2024. From Local to Global: A Graph RAG Approach to Query-Focused Summarization. arXiv:2404.16130 [cs.CL] https://arxiv.org/abs/2404.16130

    2024 arXiv

  18. [26]

    Lutfi Eren Erdogan, Nicholas Lee, Siddharth Jha, Sehoon Kim, Ryan Tabrizi, Suhong Moon, Coleman Hooper, Gopala Anumanchipalli, Kurt Keutzer, and Amir Gholami. 2024. TinyAgent: Function Calling at the Edge. arXiv:2409.00608 [cs.CL] https://arxiv.org/abs/2409.00608

    2024 arXiv

  19. [27]

    Bergstrom, Katy Börner, James A

    Santo Fortunato, Carl T. Bergstrom, Katy Börner, James A. Evans, Dirk Hel- bing, Staša Milojević, Alexander M. Petersen, Filippo Radicchi, Roberta Sinatra, Brian Uzzi, Alessandro Vespignani, Ludo Waltman, Dashun Wang, and Albert-László Barabási. 2018. Science of science.Scienc...

    2018 doi

  20. [28]

    Leo Gao, Stella Biderman, Sid Black, Laurence Golding, Travis Hoppe, Charles Foster, Jason Phang, Horace He, Anish Thite, Noa Nabeshima, Shawn Presser, and Connor Leahy. 2020. The Pile: An 800GB Dataset of Diverse Text for Language Modeling. arXiv:2101.00027 [cs.CL] https://ar...

    2020 arXiv

  21. [29]

    Tianyu Gao, Xingcheng Yao, and Danqi Chen. 2021. SimCSE: Simple Contrastive Learning of Sentence Embeddings. InConference on Empirical Methods in Natural Language Processing, Marie-Francine Moens, Xuanjing Huang, Lucia Specia, and Scott Wen-tau Yih (Eds.). Association for Comp...

    2021 doi

  22. [30]

    Yunfan Gao, Yun Xiong, Xinyu Gao, Kangxiang Jia, Jinliu Pan, Yuxi Bi, Yi Dai, Jiawei Sun, and Haofen Wang. 2023. Retrieval-augmented generation for large language models: A survey.arXiv preprint arXiv:2312.10997(2023)

    2023 arXiv

  23. [31]

    Aditi Godbole, Jabin Geevarghese George, and Smita Shandilya. 2024. Leveraging Long-Context Large Language Models for Multi-Document Understanding and Summarization in Enterprise Applications. https://doi.org/10.48550/ARXIV.2409. 18454

    2024 doi

  24. [32]

    Koustava Goswami, Lukas Lange, Jun Araki, and Heike Adel. 2023. SwitchPrompt: Learning domain-specific gated soft prompts for classification in low-resource domains.arXiv preprint arXiv:2302.06868(2023)

    2023 arXiv

  25. [34]

    Yu Gu, Robert Tinn, Hao Cheng, Michael Lucas, Naoto Usuyama, Xiaodong Liu, Tristan Naumann, Jianfeng Gao, and Hoifung Poon. 2021. Domain-Specific Language Model Pretraining for Biomedical Natural Language Processing.ACM Transactions on Computing for Healthcare3, 1 (Oct. 2021),...

    2021 doi

  26. [35]

    Edward J Hu, yelong shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2022. LoRA: Low-Rank Adaptation of Large Language Models. InInternational Conference on Learning Representations. https: //openreview.net/forum?id=nZeVKeeFYf9

    2022

  27. [36]

    Albert Q Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, De- vendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lucile Saulnier, et al . 2023. Mistral 7B.arXiv preprint arXiv:2310.06825(2023)

    2023 arXiv

  28. [37]

    Albert Q. Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, De- vendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, 11 Guillaume Lample, Lucile Saulnier, Lélio Renard Lavaud, Marie-Anne Lachaux, Pierre Stock, Teven Le Scao, Thibaut Lavril, Th...

    2023 doi arXiv

  29. [38]

    Pankaj Joshi, Aditya Gupta, Pankaj Kumar, and Manas Sisodia. 2024. Robust Multi Model RAG Pipeline For Documents Containing Text, Table & Images. In3rd International Conference on Applied Artificial Intelligence and Computing. 993–999. https://doi.org/10.1109/ICAAIC60222.2024.10574972

    2024

  30. [39]

    Vladimir Karpukhin, Barlas Oğuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen tau Yih. 2020. Dense Passage Retrieval for Open- Domain Question Answering. arXiv:2004.04906 [cs.CL] https://arxiv.org/abs/ 2004.04906

    2020 arXiv

  31. [40]

    Jacob Devlin Ming-Wei Chang Kenton and Lee Kristina Toutanova. 2019. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. In Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 4171–4186

    2019

  32. [41]

    Omar Khattab and Matei Zaharia. 2020. ColBERT: Efficient and effective passage search via contextualized late interaction over BERT. In43rd International ACM SIGIR Conference on Research and Development in Information Retrieval. 39–48

    2020

  33. [42]

    Tushar Khot, Ashish Sabharwal, and Peter Clark. 2018. SciTaiL: A textual entail- ment dataset from science question answering. InAAAI Conference on Artificial Intelligence, Vol. 32

    2018

  34. [43]

    Geewook Kim, Teakgyu Hong, Moonbin Yim, JeongYeon Nam, Jinyoung Park, Jinyeong Yim, Wonseok Hwang, Sangdoo Yun, Dongyoon Han, and Seunghyun Park. 2022. OCR-free document understanding transformer. InEuropean Confer- ence on Computer Vision. Springer, 498–517

    2022

  35. [44]

    Jakub Lála, Odhran O’Donoghue, Aleksandar Shtedritski, Sam Cox, Samuel G Rodriques, and Andrew D White. 2023. PaperQA: Retrieval-augmented generative agent for scientific research.Preprint ArXiv:2312.07559(2023)

    2023 arXiv

  36. [45]

    Esther Landhuis. 2016. Scientific literature: Information overload.Nature535 (07 2016), 457–458. https://doi.org/10.1038/nj7612-457a

    2016 doi

  37. [46]

    Le-Khac, Graham Healy, and Alan F

    Phuc H. Le-Khac, Graham Healy, and Alan F. Smeaton. 2020. Contrastive Rep- resentation Learning: A Framework and Review.CoRRabs/2010.05113 (2020). arXiv:2010.05113 https://arxiv.org/abs/2010.05113

    2020 arXiv

  38. [47]

    Jinhyuk Lee, Wonjin Yoon, Sungdong Kim, Donghyeon Kim, Sunkyu Kim, Chan Ho So, and Jaewoo Kang. 2019. BioBERT: a pre-trained biomedical language representation model for biomedical text mining.Bioinformatics36, 4 (Sept. 2019), 1234–1240. https://doi.org/10.1093/bioinformatics/btz682

    2019 doi

  39. [48]

    Patrick Lewis, Ethan Perez, Aleksandra Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Küttler, Mike Lewis, Wen-tau Yih, Tim Rocktäschel, Se- bastian Riedel, and Douwe Kiela. 2020. Retrieval-augmented generation for knowledge-intensive NLP tasks. In34th Intern...

    2020

  40. [49]

    Evans, and Lingfei Wu

    Yiling Lin, James A. Evans, and Lingfei Wu. 2022. New directions in science emerge from disconnection and discord.Journal of Informetrics16, 1 (2022), 101234. https://doi.org/10.1016/j.joi.2021.101234

    2022

  41. [50]

    Haokun Liu, Derek Tam, Mohammed Muqeeth, Jay Mohta, Tenghao Huang, Mohit Bansal, and Colin Raffel. 2022. Few-Shot Parameter-Efficient Fine-Tuning is Better and Cheaper than In-Context Learning.arXiv e-prints, Article arXiv:2205.05638 (May 2022), arXiv:2205.05638 pages. https:/...

    2022 doi arXiv

  42. [51]

    Kevin Lu, Aditya Grover, Pieter Abbeel, and Igor Mordatch. 2021. Pretrained Transformers as Universal Computation Engines. arXiv:2103.05247 [cs.LG] https://arxiv.org/abs/2103.05247

    2021 arXiv

  43. [52]

    Xinyu Ma, Jiafeng Guo, Ruqing Zhang, Yixing Fan, and Xueqi Cheng. 2022. Pre- train a Discriminative Text Encoder for Dense Retrieval via Contrastive Span Prediction. InProceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retriev...

    2022

  44. [53]

    Nicholas Matsumoto, Jay Moran, Hyunjun Choi, Miguel E Hernandez, Mythreye Venkatesan, Paul Wang, and Jason H Moore. 2024. KRAGEN: a knowledge graph- enhanced RAG framework for biomedical problem solving using large language models.Bioinformatics40, 6 (June 2024). https://doi.o...

    2024 doi

  45. [54]

    Rui Meng, Ye Liu, Shafiq Rayhan Joty, Caiming Xiong, Yingbo Zhou, and Semih Yavuz. 2024. SFR-Embedding-Mistral:Enhance Text Retrieval with Transfer Learning. Salesforce AI Research Blog. https://blog.salesforceairesearch.com/sfr- embedded-mistral/ Accessed: Apr 7, 2024

    2024

  46. [55]

    Tomas Mikolov, Kai Chen, Greg Corrado, and Jeffrey Dean. 2013. Efficient Estimation of Word Representations in Vector Space. arXiv:1301.3781 [cs.CL] https://arxiv.org/abs/1301.3781

    2013 arXiv

  47. [56]

    Radeen Mostafa, Mirza Nihal Baig, Mashaekh Tausif Ehsan, and Jakir Hasan

  48. [57]

    Mathias Wullum Nielsen and Jens Peter Andersen. 2021. Global citation inequality is on the rise.Proceedings of the National Academy of Sci- ences118, 7 (2021), e2012208118. https://doi.org/10.1073/pnas.2012208118 arXiv:https://www.pnas.org/doi/pdf/10.1073/pnas.2012208118

    2021 doi

  49. [58]

    https://doi.org/10

    G-RAG: Knowledge Expansion in Material Science. https://doi.org/10. 48550/ARXIV.2411.14592

    doi arXiv

  50. [59]

    Vik Paruchuri. [n. d.]. Texify: Tool for Converting Text to LaTeX. https://github. com/VikParuchuri/texify. Accessed: [12/7/2024]

    2024

  51. [60]

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

    2002

  52. [61]

    Malak Pirtskhalava, Anthony A Amstrong, Maia Grigolava, Mindia Chubinidze, Evgenia Alimbarashvili, Boris Vishnepolsky, Andrei Gabrielian, Alex Rosenthal, Darrell E Hurt, and Michael Tartakovsky. 2021. DBAASP v3: Database of antimi- crobial/cytotoxic activity and structure of p...

    2021

  53. [62]

    Vik Paruchuri. 2024. Marker. https://github.com/VikParuchuri/marker. Accessed: [12/7/2024]

    2024

  54. [63]

    Joseph Redmon, Santosh Divvala, Ross Girshick, and Ali Farhadi. 2016. You only look once: Unified, real-time object detection. InIEEE Conference on Computer Vision and Pattern Recognition. 779–788

    2016

  55. [64]

    Chidaksh Ravuru, Sagar Srinivas Sakhinana, and Venkataramana Runkana

  56. [65]

    arXiv:2408.14484 [cs.AI] https://arxiv.org/abs/2408.14484

    Agentic Retrieval-Augmented Generation for Time Series Analysis. arXiv:2408.14484 [cs.AI] https://arxiv.org/abs/2408.14484

    arXiv

  57. [66]

    Guobang Shi, Xinyue Kang, Fanyi Dong, Yanchao Liu, Ning Zhu, Yuxuan Hu, Hanmei Xu, Xingzhen Lao, and Heng Zheng. 2022. DRAMP 3.0: An enhanced comprehensive data repository of antimicrobial peptides.Nucleic acids research 50, D1 (2022), D488–D496

    2022

  58. [67]

    Keshav Santhanam, Omar Khattab, Jon Saad-Falcon, Christopher Potts, and Matei Zaharia. 2022. ColBERTv2: Effective and Efficient Retrieval via Lightweight Late Interaction. InConference of the North American Chapter of the Association for Computational Linguistics: Human Langua...

    2022 doi

  59. [68]

    Hunter, Costas Bekas, and Alpha A

    Philippe Schwaller, Teodoro Laino, Théophile Gaudin, Peter Bolgar, Christo- pher A. Hunter, Costas Bekas, and Alpha A. Lee. 2019. Molecular Transformer: A Model for Uncertainty-Calibrated Chemical Reaction Prediction.ACS Central Science5, 9 (Aug. 2019), 1572–1583. https://doi....

    2019 doi

  60. [69]

    Open Source. [n. d.]. The Chroma Vector Database. https://docs.trychroma.com/

  61. [70]

    Dalmeet Singh Chawla. 2022. Massive open index of scholarly papers launches. Nature(Jan. 2022). https://doi.org/10.1038/d41586-022-00138-y

    2022 doi

  62. [71]

    Peters, Abhilasha Ravichander, Kyle Richardson, Ze- jiang Shen, Emma Strubell, Nishant Subramani, Oyvind Tafjord, Pete Walsh, Luke Zettlemoyer, Noah A

    Luca Soldaini, Rodney Kinney, Akshita Bhagia, Dustin Schwenk, David Atkin- son, Russell Authur, Ben Bogin, Khyathi Chandu, Jennifer Dumas, Yanai Elazar, Valentin Hofmann, Ananya Harsh Jha, Sachin Kumar, Li Lucy, Xinxi Lyu, Nathan Lambert, Ian Magnusson, Jacob Morrison, Niklas ...

    2024 arXiv

  63. [72]

    Persson, Gerbrand Ceder, and Anubhav Jain

    Vahe Tshitoyan, John Dagdelen, Leigh Weston, Alexander Dunn, Ziqin Rong, Olga Kononova, Kristin A. Persson, Gerbrand Ceder, and Anubhav Jain. 2019. Unsupervised word embeddings capture latent knowledge from materials science literature.Nature571, 7763 (July 2019), 95–98. https...

    2019 doi

  64. [73]

    Ross Taylor, Marcin Kardas, Guillem Cucurull, Thomas Scialom, Anthony Hartshorn, Elvis Saravia, Andrew Poulton, Viktor Kerkez, and Robert Sto- jnic. 2022. Galactica: A large language model for science.arXiv preprint arXiv:2211.09085(2022)

    2022 arXiv

  65. [74]

    2024.November 2024 TOP500

    Top500. 2024.November 2024 TOP500. https://www.top500.org/lists/top500/ 2024/11/

    2024

  66. [75]

    Archit Vasan, Ozan Gokdemir, Alexander Brace, Arvind Ramanathan, Thomas Brettin, Rick Stevens, and Venkatram Vishwanath. 2024. High Performance Bind- ing Affinity Prediction with a Transformer-Based Surrogate Model. In2024 IEEE International Parallel and Distributed Processing...

    2024

  67. [76]

    UniProt Consortium. 2021. UniProt: The universal protein knowledgebase in 2021.Nucleic Acids Research49, D1 (2021), D480–D489

    2021

  68. [77]

    Richard Van Noorden. 2014. Scientists may be reaching a peak in reading habits. Nature(02 2014). https://doi.org/10.1038/nature.2014.14658

    2014

  69. [78]

    Liu, and Matt Gardner

    Johannes Welbl, Nelson F. Liu, and Matt Gardner. 2017. Crowdsourcing Mul- tiple Choice Science Questions.ArXivabs/1707.06209 (2017). https://api. 12 semanticscholar.org/CorpusID:1553193

    2017 arXiv

  70. [79]

    Chengrui Wang, Qingqing Long, Meng Xiao, Xunxin Cai, Chengjun Wu, Zhen Meng, Xuezhi Wang, and Yuanchun Zhou. 2024. BioRAG: A RAG-LLM Frame- work for Biological Question Reasoning. arXiv:2408.01107 [cs.CL] https: //arxiv.org/abs/2408.01107

    2024 arXiv

  71. [80]

    Zhe Wang and Guangshun Wang. 2004. APD: The antimicrobial peptide database. Nucleic acids research32, suppl_1 (2004), D590–D592

    2004

  72. [81]

    Lewei Yao, Runhui Huang, Lu Hou, Guansong Lu, Minzhe Niu, Hang Xu, Xi- aodan Liang, Zhenguo Li, Xin Jiang, and Chunjing Xu. 2021. FILIP: Fine- grained Interactive Language-Image Pre-Training. arXiv:2111.07783 [cs.CV] https://arxiv.org/abs/2111.07783

    2021 arXiv

  73. [82]

    Qingyun Wu, Gagan Bansal, Jieyu Zhang, Yiran Wu, Beibin Li, Erkang Zhu, Li Jiang, Xiaoyun Zhang, Shaokun Zhang, Jiale Liu, Ahmed Hassan Awadallah, Ryen W White, Doug Burger, and Chi Wang. 2023. AutoGen: Enabling Next- Gen LLM Applications via Multi-Agent Conversation. arXiv:23...

    2023 arXiv

  74. [83]

    Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, and Ming Zhou. 2020. LayoutLM: Pre-training of text and layout for document image understanding. In26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. 1192–1200

    2020

  75. [84]

    Xiaochi Zhou, Shaocong Zhang, Mehal Agarwal, Jethro Akroyd, Sebastian Mos- bach, and Markus Kraft. 2023. Marie and BERT–A Knowledge Graph Embedding Based Question Answering System for Chemistry.ACS Omega8, 36 (Aug. 2023), 33039–33057. https://doi.org/10.1021/acsomega.3c05114 13

    2023 doi

  76. [85]

    Ruochen Zhao, Hailin Chen, Weishi Wang, Fangkai Jiao, Xuan Long Do, Cheng- wei Qin, Bosheng Ding, Xiaobao Guo, Minzhi Li, Xingxuan Li, and Shafiq Joty

  77. [87]

    Xiaoqing Zheng, Haoyuan Peng, Yi Chen, Pengjing Zhang, and Wenqiang Zhang

  78. [2015]

    InTwenty- Fourth International Joint Conference on Artificial Intelligence

    Character-based parsing with convolutional neural network. InTwenty- Fourth International Joint Conference on Artificial Intelligence

  79. [2023]

    arXiv:2303.10868 [cs.CL] https://arxiv.org/abs/2303.10868

    Retrieving Multimodal Information for Augmented Generation: A Survey. arXiv:2303.10868 [cs.CL] https://arxiv.org/abs/2303.10868

    arXiv

  80. [2024]

    arXiv:2408.00798 [cs.IR] https://arxiv.org/abs/ 2408.00798

    Golden-Retriever: High-Fidelity Agentic Retrieval Augmented Generation for Industrial Knowledge Base. arXiv:2408.00798 [cs.IR] https://arxiv.org/abs/ 2408.00798

    arXiv

Pith tools

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