HiPerRAG: High-Performance Retrieval Augmented Generation for Scientific Insights
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908353651802112 |
|---|---|
| author | Gokdemir, Ozan Siebenschuh, Carlo Brace, Alexander Wells, Azton Hsu, Brian Hippe, Kyle Setty, Priyanka V. Ajith, Aswathy Pauloski, J. Gregory Sastry, Varuni Foreman, Sam Zheng, Huihuo Ma, Heng Kale, Bharat Chia, Nicholas Gibbs, Thomas Papka, Michael E. Brettin, Thomas Alexander, Francis J. Anandkumar, Anima Foster, Ian Stevens, Rick Vishwanath, Venkatram Ramanathan, Arvind |
| author_facet | Gokdemir, Ozan Siebenschuh, Carlo Brace, Alexander Wells, Azton Hsu, Brian Hippe, Kyle Setty, Priyanka V. Ajith, Aswathy Pauloski, J. Gregory Sastry, Varuni Foreman, Sam Zheng, Huihuo Ma, Heng Kale, Bharat Chia, Nicholas Gibbs, Thomas Papka, Michael E. Brettin, Thomas Alexander, Francis J. Anandkumar, Anima Foster, Ian Stevens, Rick Vishwanath, Venkatram Ramanathan, Arvind |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_04846 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | HiPerRAG: High-Performance Retrieval Augmented Generation for Scientific Insights Gokdemir, Ozan Siebenschuh, Carlo Brace, Alexander Wells, Azton Hsu, Brian Hippe, Kyle Setty, Priyanka V. Ajith, Aswathy Pauloski, J. Gregory Sastry, Varuni Foreman, Sam Zheng, Huihuo Ma, Heng Kale, Bharat Chia, Nicholas Gibbs, Thomas Papka, Michael E. Brettin, Thomas Alexander, Francis J. Anandkumar, Anima Foster, Ian Stevens, Rick Vishwanath, Venkatram Ramanathan, Arvind Information Retrieval Computational Engineering, Finance, and Science Computation and Language Distributed, Parallel, and Cluster Computing Machine Learning H.3.3; I.2.7 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. |
| title | HiPerRAG: High-Performance Retrieval Augmented Generation for Scientific Insights |
| topic | Information Retrieval Computational Engineering, Finance, and Science Computation and Language Distributed, Parallel, and Cluster Computing Machine Learning H.3.3; I.2.7 |
| url | https://arxiv.org/abs/2505.04846 |