arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2505.08041q-bio.BM
收藏

Protein FID:改进蛋白质结构生成模型的评估

Protein FID: Improved Evaluation of Protein Structure Generative Models

Felix Faltings, Hannes Stark, Tommi Jaakkola, Regina Barzilay

AI总结:

本文提出 Protein FID,通过潜空间分布相似性评估蛋白质结构生成模型,验证其对扰动、结构聚类和层级关系的判别能力,并发现现有模型仍未充分拟合 PDB 蛋白质分布。

摘要:

蛋白质结构生成模型近期引发了广泛兴趣,但如何通过计算对其进行有意义的评估仍是一个活跃的研究领域。尽管现有指标推动了有益进展,但它们无法反映模型对训练数据所代表设计空间的采样效果。我们主张采用蛋白质 Fréchet Inception Distance(FID)指标,在语义有意义的潜空间中度量分布相似性,以补充当前评估。我们的 FID 在蛋白质结构扰动下表现出理想特性,并能正确重现蛋白质样本之间的相似性:它与最优传输距离相关,并能恢复 FoldSeek 聚类和 CATH 层级结构。使用 FID 评估当前蛋白质结构生成模型表明,这些模型尚不足以建模 PDB 蛋白质的分布。

英文摘要:

Protein structure generative models have seen a recent surge of interest, but meaningfully evaluating them computationally is an active area of research. While current metrics have driven useful progress, they do not capture how well models sample the design space represented by the training data. We argue for a protein Frechet Inception Distance (FID) metric to supplement current evaluations with a measure of distributional similarity in a semantically meaningful latent space. Our FID behaves desirably under protein structure perturbations and correctly recapitulates similarities between protein samples: it correlates with optimal transport distances and recovers FoldSeek clusters and the CATH hierarchy. Evaluating current protein structure generative models with FID shows that they fall short of modeling the distribution of PDB proteins.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑