Unverified50% confidenceFactExact time
GNoME (Graph Networks for Materials Exploration), released by DeepMind, predicts crystal structure stability through graph neural networks and screened out millions of potential new material candidates at once.
1
Sources
50%
Confidence
Long-term
Relevance
9/16/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedDeepMind's GNoME project discovered 2.2 million new crystal structures at once, equivalent to 800 years of accumulated human materials science.71% similarUnverifiedDeepMind's GraphCast uses graph neural networks to capture the irregular grid structure of the atmospheric layer.65% similarUnverifiedDeepMind、Meta、Mistral等实验室已在多个基准测试中证明，高质量合成数据可以大幅减少对人工标注的依赖62% similarUnverifiedThe core innovation of GraphSAGE lies in sampling neighbors combined with a learnable aggregation function, enabling GNNs to scale to large graphs and support inductive learning.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/926111API
curl https://kongchang.com/api/v1/knowledge/claims/926111MCP
get_claim(id=926111)