Unverified50% confidenceFactExact time
With only 1000 molecules of training data, randomized SAFE representations increase the yield of genuinely novel molecules by more than 450x.
1
Sources
50%
Confidence
Long-term
Relevance
10/30/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedRandomizing only the fragment order of SAFE achieves just one-seventh of the genuinely novel molecule yield of simultaneous dual-axis randomization.76% similarUnverifiedSingle-token output can theoretically achieve a 50x throughput improvement compared to standard generation (assuming 50 decoding steps are required).67% similarUnverified在同一张英伟达H100显卡上，DiffusionGemma每秒可生成约1500个token，而上一代自回归模型仅能生成约300个，快了近5倍64% similarUnverified研究表明即使架构和超参数完全相同，不同随机种子训练出的模型可能在准确率上产生 0.5% 到 2% 的自然波动63% similarUnverifiedIn real physical laboratory validation of molecular design, the model achieved a hit rate of nearly 50% across 12 targets, compared to the industry average of approximately 10% to 15%.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1114330API
curl https://kongchang.com/api/v1/knowledge/claims/1114330MCP
get_claim(id=1114330)