arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2602.05285cs.LG
收藏

通过嵌入优化实现蛋白质扩散模型推理时间的鲁棒引导

Robust Inference-Time Steering of Protein Diffusion Models via Embedding Optimization

Flatiron Institute(Flatiron研究所)

Minhuan Li, Jiequn Han, Pilar Cossio, Luhuan Wu

AI总结:

本文提出EmbedOpt框架,通过优化嵌入空间使蛋白质结构先验与实验约束对齐,在稀疏距离约束和冷冻电镜图匹配中表现更优,且对超参数更鲁棒。

摘要:

结构生物物理中的核心挑战是生成既物理合理又符合实验测量的生物分子构象。尽管序列到结构扩散模型提供强大先验,但后验采样方法通过梯度扰动原子坐标来引导生成,当目标位于先验低密度区域时,需要强烈加权似然,这可能破坏采样并敏感于超参数。本文提出EmbedOpt,一种推理时间引导框架,引入正交优化轴:而非在固定先验下进行后验采样,EmbedOpt直接优化先验,通过更新模型的条件嵌入。该嵌入空间编码丰富的共进化信号,优化它会将结构先验调整以符合实验约束。实验证明,EmbedOpt在稀疏距离约束上与基于坐标的后验采样基线匹配,在冷冻电镜图匹配中表现更优,包括真实、噪声实验数据。此外,EmbedOpt的平滑优化行为使其在超参数跨度两个数量级时仍具鲁棒性,并能用更少扩散步骤获得可比性能。代码可在https://github.com/rs-station/embedopt获取。

英文摘要:

A core challenge in structural biophysics is generating biomolecular conformations that are both physically plausible and consistent with experimental measurements. While sequence-to-structure diffusion models provide powerful priors, posterior sampling methods steer generation by perturbing atomic coordinates with gradients from experimental likelihoods. However, when the target lies in a low-density region of the prior, these methods require aggressive upweighting of the likelihood that can destabilize sampling and be sensitive to hyperparameters. We propose EmbedOpt, an inference-time steering framework that introduces an orthogonal optimization axis: rather than performing posterior sampling under a fixed prior, EmbedOpt directly optimizes the prior by updating the model's conditional embedding. This embedding space encodes rich coevolutionary signals, so optimizing it shifts the structural prior to align with experimental constraints. Empirically, EmbedOpt matches coordinate-based posterior sampling baselines on sparse distance constraints and outperforms them on cryo-electron microscopy map fitting, including real, noisy experimental ones. Furthermore, EmbedOpt's smooth optimization behavior yields robustness to hyperparameters spanning two orders of magnitude and enables comparable performance with fewer diffusion steps. Code is available at https://github.com/rs-station/embedopt.

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。

↑