De novo Generation for Molecular Structure Elucidation from Mass Spectrometry
Microsoft Research13033 字 (约 53 分钟)
75
Microsoft researchers propose Diffs, a novel generative model using diffusion models and graph neural networks to infer molecular structures from mass spectrometry data, significantly improving accuracy and efficiency.
入选理由：Diffs 模型结合了扩散模型和图神经网络，能够从质谱数据中生成分子结构，准确率达到 90% 以上。
FeaturedVideo#AI for Science#Generative Models#Mass Spectrometry#Molecular Structure Elucidation#Diffs ModelEnglish