研究人员引入了超维度指纹（HDF）作为一种新颖的分子表示方法，旨在克服传统基于哈希的指纹固有的信息丢失问题。与图神经网络不同，HDF 利用高维向量上的代数运算，无需进行特定任务的训练即可降低计算需求。该新方法在各种属性预测基准测试中表现出卓越的性能和一致性，比传统方法更能有效地保留分子相似性，尤其是在较低维度下。 AI
影响 提供了一种无需训练、更具表现力的传统分子指纹替代方案，有望提高药物发现和材料科学的效率。
排序理由 介绍分子表示新计算方法的学术论文。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2604.27810v1 Announce Type: new Abstract: Computational molecular representations underpin virtual screening, property prediction, and materials discovery. Conventional fingerprints are efficient and deterministic but lose structural information through hash-based compressi…
Computational molecular representations underpin virtual screening, property prediction, and materials discovery. Conventional fingerprints are efficient and deterministic but lose structural information through hash-based compression, particularly at low dimensionalities. Learne…
Computational molecular representations underpin virtual screening, property prediction, and materials discovery. Conventional fingerprints are efficient and deterministic but lose structural information through hash-based compression, particularly at low dimensionalities. Learne…