Michael Plainer
@plainer.bsky.social
PhD student @ ELIZA TU/FU Berlin - plainer.dev
(5/n) With this, we can run coarse-grained Langevin dynamics directly, without the need for any priors or force labels.
This works across biomolecular systems including fast-folding proteins like Chignolin and BBA.
Here is a comparison with and without our regularization:
This works across biomolecular systems including fast-folding proteins like Chignolin and BBA.
Here is a comparison with and without our regularization:
November 6, 2025 at 2:41 PM
(5/n) With this, we can run coarse-grained Langevin dynamics directly, without the need for any priors or force labels.
This works across biomolecular systems including fast-folding proteins like Chignolin and BBA.
Here is a comparison with and without our regularization:
This works across biomolecular systems including fast-folding proteins like Chignolin and BBA.
Here is a comparison with and without our regularization:
(4/n) Our solution:
We train an energy-based diffusion model and regularize it to satisfy the Fokker–Planck equation.
This enforces consistency between:
- The density recovered via denoising
- The potential energy learned at t = 0
Result: the same model can be used for sampling AND simulation.
We train an energy-based diffusion model and regularize it to satisfy the Fokker–Planck equation.
This enforces consistency between:
- The density recovered via denoising
- The potential energy learned at t = 0
Result: the same model can be used for sampling AND simulation.
November 6, 2025 at 2:41 PM
(4/n) Our solution:
We train an energy-based diffusion model and regularize it to satisfy the Fokker–Planck equation.
This enforces consistency between:
- The density recovered via denoising
- The potential energy learned at t = 0
Result: the same model can be used for sampling AND simulation.
We train an energy-based diffusion model and regularize it to satisfy the Fokker–Planck equation.
This enforces consistency between:
- The density recovered via denoising
- The potential energy learned at t = 0
Result: the same model can be used for sampling AND simulation.
(2/n) The problem: classical diffusion models learn scores that reproduce equilibrium samples, but the corresponding energy-based parameterization is not consistent.
So if you try to use the learned energy to derive forces, the dynamics are wrong, even if the samples themselves look fine.
So if you try to use the learned energy to derive forces, the dynamics are wrong, even if the samples themselves look fine.
November 6, 2025 at 2:41 PM
(2/n) The problem: classical diffusion models learn scores that reproduce equilibrium samples, but the corresponding energy-based parameterization is not consistent.
So if you try to use the learned energy to derive forces, the dynamics are wrong, even if the samples themselves look fine.
So if you try to use the learned energy to derive forces, the dynamics are wrong, even if the samples themselves look fine.
(1/n) Can diffusion models simulate molecular dynamics instead of just generating independent samples?
In our NeurIPS 2025 paper, we train energy-based diffusion models that can do both:
- Generate independent samples
- Learn the underlying potential 𝑼
🧵👇
Paper: arxiv.org/abs/2506.17139
In our NeurIPS 2025 paper, we train energy-based diffusion models that can do both:
- Generate independent samples
- Learn the underlying potential 𝑼
🧵👇
Paper: arxiv.org/abs/2506.17139
November 6, 2025 at 2:41 PM
(1/n) Can diffusion models simulate molecular dynamics instead of just generating independent samples?
In our NeurIPS 2025 paper, we train energy-based diffusion models that can do both:
- Generate independent samples
- Learn the underlying potential 𝑼
🧵👇
Paper: arxiv.org/abs/2506.17139
In our NeurIPS 2025 paper, we train energy-based diffusion models that can do both:
- Generate independent samples
- Learn the underlying potential 𝑼
🧵👇
Paper: arxiv.org/abs/2506.17139