repoGitHubTrust 82 · PrimaryPublished 4h agoLive · 31m ago
janosh/matbench-discovery
An evaluation framework for machine learning models simulating high-throughput materials discovery.
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Implements
paperBeyond Drug Discovery: The Nanotechnology Molecular Optimization (NMO) BenchmarkpaperGrounded autonomous research: a fault-tolerant LLM pipeline from corpus to manuscript in frontier computational physicspaperBeyond Adam: SOAP and Muon for Faster, Label-Efficient Training of Machine Learning Interatomic PotentialspaperGraph-Native Reinforcement Learning Enables Traceable Scientific Hypothesis Generation through Conceptual Recombination
Related to
Related across the graph
paperBeyond Drug Discovery: The Nanotechnology Molecular Optimization (NMO) BenchmarkpaperBeyond Adam: SOAP and Muon for Faster, Label-Efficient Training of Machine Learning Interatomic PotentialscompanyLattice LabspaperGrounded autonomous research: a fault-tolerant LLM pipeline from corpus to manuscript in frontier computational physicspaperGraph-Native Reinforcement Learning Enables Traceable Scientific Hypothesis Generation through Conceptual Recombination