Summary
Stronger diffusion guidance reduces diversity of generated outputs (1–2). This was observed in sequence generation using low-N fitness data. In Feynman-Kac steering, increasing the reward weight is equivalent to deprioritizing the prior distribution, and concentrates particle weights; this can collapse the sampled distribution, while low recovers behavior closer to the prior.
Figures
Ref (1)
1.
Yang J, Chu W, Khalil D, Astudillo R, Wittmann BJ, Arnold FH, et al. Steering Generative Models with Experimental Data for Protein Fitness Optimization. 2025; Available from: https://arxiv.org/abs/2505.15093
2.
Singhal R, Horvitz Z, Teehan R, Ren M, Yu Z, McKeown K, et al. A General Framework for Inference-time Scaling and Steering of Diffusion Models. In: Proceedings of the 42nd International Conference on Machine Learning. PMLR; 2025. p. 55810–27. Available from: https://proceedings.mlr.press/v267/singhal25b.html
3.
Hartman E, Wallin J, Malmström J, Olsson J. Controllable protein design with particle-based Feynman-Kac steering. 2025; Available from: https://arxiv.org/abs/2511.09216