ArtifactLinker uses GNNs and LLMs to find SOTA models automatically

Researchers have developed ArtifactLinker, a novel framework designed to automatically discover state-of-the-art (SOTA) models for specific datasets by analyzing the vast landscape of scientific artifacts. The system models platforms like Hugging Face as artifact graphs, utilizing graph neural networks or large language models to predict promising model-dataset links. These predictions are then verified through LLM-based agents that conduct coding experiments, with a new benchmark called ArtifactBench comprising over 14,000 artifacts and 51,000 relations to evaluate the framework's effectiveness. AI

IMPACT Automates the discovery of SOTA models, potentially accelerating research and development cycles.

RANK_REASON The cluster contains a research paper detailing a new framework and benchmark for discovering state-of-the-art models. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• alphaXiv
• ArtifactBench
• ArtifactLinker
• arXiv
• CatalyzeX Code Finder for Papers
• DagsHub
• Gotit.pub
• graph neural networks
• Haofei Yu
• Hugging Face
• large-language models
• ScienceCast

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →