betaFeedback

WorksComputer Science

Kolmogorov-Chapman Maps: Trajectory-Free Learning for Simulating Molecular Dynamics

Abstract

Stochastic differential equations (SDEs) describe stochastic processes across a wide range of scientific domains, including physics, biology, and finance. Although individual numerical integration steps are inexpensive, observing meaningful events or transitions can require millions of sequential steps, resulting in substantial wall-clock simulation time and poor hardware utilization. Deep learning offers an alternative by amortizing many integration steps into a single neural network forward pass, potentially enabling much faster simulation. However, existing supervised approaches typically require long simulated trajectory data, which restricts their applicability. In this work, we introduce Kolmogorov-Chapman Maps (KC-Maps), a method for learning the transition kernel of an SDE that relies only on the trajectory-free marginal samples together with the known drift and diffusion coefficients. Given an initial time and state of a process, KC-Maps models its distribution integrated to any given time, corresponding to the weak solution of the SDE. To recover the correct dynamics, KC-Maps satisfies two key properties: (i) it satisfies the boundary conditions, which we incorporate directly into the neural network architecture, and (ii) it respects the Kolmogorov–Chapman equation, which provides an efficient training objective. We empirically demonstrate that KC-Maps learns to integrate the dynamics of peptide molecular systems and outperforms classical molecular dynamics in terms of GPU-hours.

Introduction

Stochastic differential equations (SDEs) provide a fundamental framework for describing stochastic processes, with applications across physics, finance, and biology. A prominent example is molecular dynamics ([34, 20]), where Langevin equations describe the evolution of the atomic coordinates and momenta under interatomic forces and random fluctuations. Classical approaches, such as Euler-Maruyama ([29]) and Langevin splitting schemes ([33]), approximate the dynamics through a sequence of short-time updates and provide systematic control over discretization error. However, such methods are inherently sequential, with computational cost scaling with the number of integration steps. In molecular dynamics, femtosecond-scale time steps (10−1510^{-15} seconds) are typically required to maintain numerical stability and resolve fast motions such as bond vibrations, while conformational changes and transitions between metastable states can occur over orders-of-magnitude longer timescales, e.g. 10−610^{-6} seconds. Reaching these slow phenomena can therefore require an enormous number of sequential force evaluations, making long-horizon molecular simulation computationally expensive.

Recently, numerous approaches were developed to learn surrogates that can accelerate molecular simulation by predicting the evolution over larger time intervals. Generative modeling approaches ([27, 51]) learn conditional distributions of future molecular configurations from simulated trajectories. Other approaches ([4]) learn direct deterministic updates of the positions and momenta. However, both formulations learn these transitions from pairs of states extracted from reference trajectories. Generating these pairs still requires numerical integration with time steps typically of the order of 10−1510^{-15} s, incurring substantial computational cost to produce the training data. Moreover, equilibrium samples can also be obtained using methods such as sequential Monte Carlo [41] or parallel tempering [21]. Yet, their sampling transitions do not necessarily follow the physical dynamics of interest, so the resulting data may not provide the time-dependent pairs required by supervised methods.

The underlying SDE, however, provides a different source of supervision. Although the finite-time transition kernel is generally unavailable in closed form, the drift and diffusion specify the dynamics locally. This raises the possibility of learning finite-time transitions from the governing SDE itself, without first generating reference trajectories. Recent approaches in flow-based generative modeling [16, 52, 36, 3] use consistency relations to connect the underlying vector fields with the finite-time maps [53, 13, 6], enabling accelerated generative sampling. Hamiltonian Flow Maps [49] apply this consistency principle to molecular dynamics, learning Hamiltonian evolution from independent states and instantaneous force fields without trajectory supervision. For stochastic dynamics, we seek an analogous formulation at the level of conditional distributions, connecting their finite-time evolution to the known drift and diffusion of the SDE.

In this work, we present Kolmogorov-Chapman Maps (KC-maps), a trajectory-free method for learning finite-time transitions of SDEs. Our method uses uncorrelated samples of the initial state together with the known vector field and diffusion, without requiring trajectory supervision or a prescribed probability path. We derive a distributional consistency objective from the Kolmogorov–Chapman equation, enforcing agreement between transitions composed over different time scales; for example, two successive 5 fs transitions should agree in distribution with a single 10 fs transition. To enable tractable density evaluation and satisfy SDE boundary conditions, we parameterize transitions densities using a conditional normalizing flow with a Gaussian reference kernel. The parameterization is constrained to recover the appropriate short-time behavior of the underlying SDE, while additional design choices enable learning and sampling over substantially longer time intervals.

The formulation of KC-maps applies to general SDEs. Here, we focus on molecular Langevin dynamics because of both its computational difficulty and its importance to applications such as drug discovery. We first conduct an empirical study on a small molecular system to identify important design choices required for stable and accurate learning, including the training objective, design of the normalizing flow, and stabilization techniques for sampling over long time intervals. We then apply KC-maps to larger scale molecular systems and highlight the following main results:

  • Using only equilibrium samples and force-field evaluations, without trajectory supervision, we learn transition distributions over intervals of up to 100 fs for alanine dipeptide (AL2), tetrapeptide (AL4), and hexapeptide (AL6), replacing up to 100 sequential MD integration steps with a single learned transition.

  • We demonstrate higher simulation throughput and faster exploration of conformational modes. On AL6, our method achieves approximately 2.3× aggregate simulation throughput (ns per GPU-hour) of MD on a single L40s GPU and reaches three of the four TICA modes while MD baseline still misses two modes after an hour. Additionally, our method is able to recover all four modes of AL6 with longer simulations despite the training data only covering three of the four modes.

  • Beyond equilibrium sampling, KC-maps captures molecular kinetics, with mean first passage times (MFPT), mode autocorrelations, and transition probabilities closely matching ground-truth MD.

Background

Stochastic Differential Equations. Stochastic differential equations (SDEs) describe the stochastic process XtX_t via the differential equation for their time-increments dXtdX_t, i.e.

dXt=vt(Xt) dt+gt dWt,(1)dX_t = v_t(X_t)\,dt + g_t\,dW_t, \tag*{(1)}

where gt∈Rd×dg_t \in\mathbb{R}^{d \times d} is a state-independent diffusion matrix, and WtW_t is standard Brownian motion in Rd\mathbb{R}^d. We will refer to the solution of the Cauchy problem of this SDE as a transition kernel, i.e. given the initial condition Xs=xX_s = x, we denote the density of the distribution of XtX_t as ps,t(y∣x):=p(Xt=y∣Xs=x)p_{s,t}(y \mid x) := p(X_t = y \mid X_s = x).

Figure 1. A single KC-maps rollout explores the landscape of metastable alanine tetrapeptide conformers. Starting from a single conformer X0X_0 (left), KC-maps is rolled out for 200 ns. Orange arrows trace the first visit to each newly discovered basin. Representative conformers from each basin are shown on the right, colored by basin identity, along with their backbone torsion angles (Φ2,Φ3)(\Phi_2, \Phi_3).

Molecular Dynamics. The formalism of SDEs is especially useful for describing the evolution of molecular systems. Their state is described by the position XtX_t and momentum PtP_t and its time-evolution is modeled by the second-order dynamics, which we will write in the matrix form

\begin{bmatrix} dX_t \\ dP_t \end{bmatrix} = \begin{bmatrix} M^{-1}P_t \\ v_t(X_t)-\gamma P_t \end{bmatrix}dt + \begin{bmatrix} 0 & 0 \\ 0 & g_t \end{bmatrix}dW_t, \tag*{(2)}\end{bmatrix}

where vtv_t is the force field driving the system, γ\gamma is the friction coefficient, MM is the diagonal mass matrix, and gtg_t is the diffusion coefficient that depends on the temperature of the system. For vt(X)=−∇U(X)v_t(X)=-\nabla U(X) where U(X)U(X) is a scalar potential and gt=2γkBTMg_t=\sqrt{2\gamma k_B T M}, the stationary distribution of XtX_t is the Boltzmann distribution, i.e. the equilibrium of the molecular system. Thus, in practice, simulating this SDE for long enough time one can sample from the Boltzmann distribution describing different states — conformations — of the system.

Normalizing Flows. Generative modeling of the empirical data has two principal components: maximizing the model likelihood on the observed data and generating new samples under the learned density model. Normalizing flows satisfy both criteria by parameterizing bijective maps with tractable Jacobians. The building block of normalizing flows is the affine coupling layer Ψk\Psi_k, which transforms the state x(k)=[x1(k),x2(k)]x^{(k)}=[x_1^{(k)},x_2^{(k)}] into x(k+1)=Ψk(x(k))x^{(k+1)}=\Psi_k(x^{(k)}) as

x1(k+1)=x1(k),(3)x_1^{(k+1)}=x_1^{(k)}, \tag*{(3)}
x2(k+1)=x2(k)⊙exp⁡(ak(x1(k);θ,c))+bk(x1(k);θ,c),(4)x_2^{(k+1)}=x_2^{(k)}\odot\exp\left(a_k(x_1^{(k)};\theta,c)\right)+b_k(x_1^{(k)};\theta,c), \tag*{(4)}

where ak(⋅;θ,c):Rd1→Rd2a_k(\cdot;\theta,c):\mathbb{R}^{d_1}\to\mathbb{R}^{d_2} and bk(⋅;θ,c):Rd1→Rd2b_k(\cdot;\theta,c):\mathbb{R}^{d_1}\to\mathbb{R}^{d_2} are functions parameterized by the neural networks with parameters θ\theta and, optionally, taking a condition cc as an input. This layer is invertible, i.e. one can find [x1(k),x2(k)][x_1^{(k)},x_2^{(k)}] from [x1(k+1),x2(k+1)][x_1^{(k+1)},x_2^{(k+1)}], and the Jacobian of this transformation is available in a closed-form without additional computations

log⁡det⁡∣∂Ψk(x(k))∂x(k)∣=log⁡[I0∂x2(k+1)∂x1(k)Iexp⁡(ak(x1(k);θ,c))]=∑iak(x1(k);θ,c)i.(5)\log\det\left|\frac{\partial\Psi_k(x^{(k)})}{\partial x^{(k)}}\right| = \log\left[ \begin{array}{cc} I & 0 \\ \frac{\partial x_2^{(k+1)}}{\partial x_1^{(k)}} & I\exp\left(a_k(x_1^{(k)};\theta,c)\right) \end{array} \right] = \sum_i a_k(x_1^{(k)};\theta,c)_i. \tag*{(5)}

Finally, normalizing flows parameterize a density p(x∣θ,c)p(x\mid\theta,c) as a push-forward map of some simple density pref(x)p^{\mathrm{ref}}(x) through the composition of affine layers Ψ=ΨK∘…∘Ψ1\Psi=\Psi_K\circ\ldots\circ\Psi_1

log⁡p(x∣θ,c)=log⁡pref(Ψ−1(x;θ,c))−∑klog⁡det⁡∣∂Ψk(x(k))∂x(k)∣.(6)\log p(x\mid\theta,c)=\log p^{\mathrm{ref}}\left(\Psi^{-1}(x;\theta,c)\right)-\sum_k\log\det\left|\frac{\partial\Psi_k(x^{(k)})}{\partial x^{(k)}}\right|. \tag*{(6)}

Kolmogorov-Chapman Maps

Our method is designed around the classical definition of a diffusion process as a Markov process with continuous trajectories and infinitesimal first and second moments prescribed by the drift and diffusion coefficient [32, 12]. The process defined this way is a weak solution of the corresponding SDE [54]. Therefore, our parametric model has to satisfy two conditions: (i) parameterize a valid Markov process, i.e. satisfy the Kolmogorov–Chapman equation, which we enforce through training (ii) the short-time behavior has to match the corresponding SDE, i.e. commit no jumps and have same first two moments, which we incorporate directly into our parameterization. We now proceed to formulating this statement formally.

For the SDE defined in Eq. (1), the infinitesimal change of the density ps,t(y∣x)p_{s,t}(y \mid x) is defined by the following boundary conditions. ∀ε>0\forall\varepsilon> 0, we have

lim⁡τ→01τ∫∥y−x∥>εdy ps,s+τ(y∣x)=0,(7)\lim_{\tau\to0} \frac{1}{\tau} \int_{\lVert y-x\rVert> \varepsilon} \mathrm{d}y \, p_{s,s+\tau}(y \mid x) = 0, \tag*{(7)}

(delta function with no jumps)

lim⁡τ→01τ∫∥y−x∥≤εdy ps,s+τ(y∣x)(y−x)=vs(x),(8)\lim_{\tau\to0} \frac{1}{\tau} \int_{\lVert y-x\rVert\le\varepsilon} \mathrm{d}y \, p_{s,s+\tau}(y \mid x)(y-x) = v_s(x), \tag*{(8)}

(matching first moment)

lim⁡τ→01τ∫∥y−x∥≤εdy ps,s+τ(y∣x)(y−x)(y−x)⊤=gsgs⊤,(9)\lim_{\tau\to0} \frac{1}{\tau} \int_{\lVert y-x\rVert\le\varepsilon} \mathrm{d}y \, p_{s,s+\tau}(y \mid x)(y-x)(y-x)^{\top} = g_s g_s^{\top}, \tag*{(9)}

(matching second moment)

i.e. for t→st \to s, the density converges to the delta-function and makes no jumps, the change of the first moment equals the vector field and the change of the second moment equals the diffusion coefficient. Furthermore, any Markov transition kernel satisfies the Kolmogorov–Chapman equation

ps,t(y∣x)=∫dx′ pr,t(y∣x′)ps,r(x′∣x).(10)p_{s,t}(y \mid x) = \int\mathrm{d}x' \, p_{r,t}(y \mid x')p_{s,r}(x' \mid x). \tag*{(10)}

This equation is the stochastic analog of the semi-group consistency condition in the ODEs, which is used in the Flow Maps literature ([6, 59]).

These two conditions define the SDE Eq. (1) and vice versa, the weak solution of an SDE satisfies both these conditions. The first statement we prove in Section A.1 using the classical result of probability theory ([54]), the second statement is a textbook statement.

Proposition 1 (Identification of the transition kernel). Assume that gsg_s is continuous in ss and that vs(x)v_s(x) is continuous, locally Lipschitz in xx, and of at most linear growth, uniformly for ss in compacts. Let ps,t(y∣x)p_{s,t}(y \mid x) be a measurable family of nonnegative, normalized conditional densities. Suppose it satisfies the boundary conditions in Eqs. (7) to (9), with limits locally uniform in (s,x)(s,x), and the Kolmogorov–Chapman equation in Eq. (10). Then ps,t(y∣x)p_{s,t}(y \mid x) coincides (almost everywhere) with the transition density of the SDE in Eq. (1). Conversely, the SDE’s transition kernel satisfies both the boundary conditions and Eq. (10).

Kolmogorov–Chapman Constraints (Our Objective)

We parameterize our transition kernel qs,t(y∣x;θ)q_{s,t}(y \mid x;\theta) such that the boundary conditions are satisfied (see the following section; Section B.2 details the case for learning the drift instead), which leaves the Kolmogorov–Chapman equation as the sole optimization objective. The most straightforward way to train the model would be to minimize the discrepancy between both sides of the Kolmogorov–Chapman equation for the parameterized model, i.e.

LKC(θ)=DKL(∫dx′ qr,t(y∣x′;θ)qs,r(x′∣x;θ) ∥ qs,t(y∣x;θ)), ∀x,t,s,r.(11)\mathcal{L}_{\mathrm{KC}}(\theta) = D_{\mathrm{KL}}\left(\int\mathrm{d}x' \, q_{r,t}(y \mid x';\theta)q_{s,r}(x' \mid x;\theta) \,\middle\|\, q_{s,t}(y \mid x;\theta)\right), \ \forall x,t,s,r. \tag*{(11)}

However, as written, this function does not have an unbiased estimator of the gradient. Therefore, we propose to use the “stop-gradient trick” ubiquitously employed in the generative modeling literature ([13, 6]). Setting η=stopgrad⁡(θ)\eta= \operatorname{stopgrad}(\theta) as the parameters in the first argument of the KL-divergence, we get (see Section B.1)

References

  1. [1]Tara Akhound-Sadegh, Jarrid Rector-Brooks, Avishek Joey Bose, Sarthak Mittal, Pablo Lemos, Cheng-Hao Liu, Marcin Sendera, Siamak Ravanbakhsh, Gauthier Gidel, Yoshua Bengio, Nikolay Malkin, and Alexander Tong. Iterated Denoising Energy Matching for Sampling from Boltzmann Densities, February 2024. URL http://arxiv.org/abs/2402.06121. arXiv:2402.06121 [cs, stat].
  2. [2]Tara Akhound-Sadegh, Jungyoon Lee, Avishek Joey Bose, Valentin De Bortoli, Arnaud Doucet, Michael M. Bronstein, Dominique Beaini, Siamak Ravanbakhsh, Kirill Neklyudov, and Alexander Tong. Progressive inference-time annealing of diffusion models for sampling from boltzmann densities, 2025. URL https://arxiv.org/abs/2506.16471.
  3. [3]Michael S. Albergo and Eric Vanden-Eijnden. Building Normalizing Flows with Stochastic Interpolants, March 2023. URL http://arxiv.org/abs/2209.15571. arXiv:2209.15571 [cs].
  4. [4]Filippo Bigi, Sanggyu Chong, Agustinus Kristiadi, and Michele Ceriotti. Flashmd: long-stride, universal prediction of molecular dynamics, 2026. URL https://arxiv.org/abs/2505.19350.
  5. [5]Boffi, N. M., Albergo, M. S., & Vanden-Eijnden, E. (2024). Flow Map Matching. http://arxiv.org/abs/2406.07507
  6. [6]Nicholas M. Boffi, Michael S. Albergo, and Eric Vanden-Eijnden. How to build a consistency model: Learning flow maps via self-distillation, October 2025. URL https://arxiv.org/abs/2505.18825. arXiv:2505.18825 [cs].
  7. [7]Bradbury, J., Frostig, R., Hawkins, P., Johnson, M. J., Katariya, Y., Leary, C., Maclaurin, D., Necula, G., Paszke, A., VanderPlas, J., Wanderman-Milne, S., & Zhang, Q. (2018). JAX: composable transformations of Python+NumPy programs. http://github.com/jax-ml/jaxgithub.com/jax-ml/jax
  8. [9]Giovanni Bussi, Davide Donadio, and Michele Parrinello. Canonical sampling through velocity rescaling. The Journal of Chemical Physics, 126(1):014101, 2007.pubmed.ncbi.nlm.nih.gov/17212484
  9. [10]Purushottam D. Dixit and Ken A. Dill. Inferring microscopic kinetic rates from stationary state distributions. Journal of Chemical Theory and Computation, 10(8):3002–3005, 2014. doi: 10.1021/ct5001389.DOI
  10. [11]Eastman, P., Swails, J., Chodera, J. D., McGibbon, R. T., Zhao, Y., Beauchamp, K. A., Wang, L.-P., Simmonett, A. C., Harrigan, M. P., Stern, C. D., & al., . et . (2017). OpenMM 7: Rapid development of high performance algorithms for molecular dynamics. PLoS Computational Biology, 13(7), e1005659.
  11. [12]Willy Feller. Zur theorie der stochastischen prozesse: Existenz-und eindeutigkeitssätze. Mathematische Annalen, 113(1):113–160, 1937.DOI
  12. [13]Zhengyang Geng, Mingyang Deng, Xingjian Bai, J. Zico Kolter, and Kaiming He. Mean Flows for One-step Generative Modeling, May 2025. URL http://arxiv.org/abs/2505.13447. arXiv:2505.13447 [cs].
  13. [14]Alan Grossfield, Paul N. Patrone, Daniel R. Roe, Andrew J. Schultz, Daniel W. Siderius, and Daniel M. Zuckerman. Best practices for quantification of uncertainty and sampling quality in molecular simulations [article v1.0]. Living Journal of Computational Molecular Science, 1(1):5067, 2018. doi: 10.33011/livecoms.1.1.5067.DOI
  14. [15]W. K. Hastings. Monte Carlo sampling methods using Markov chains and their applications. Biometrika, 57(1):97–109, 1970. doi: 10.1093/biomet/57.1.97. URL https://doi.org/10.1093/biomet/57.1.97.
  15. [16]Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising Diffusion Probabilistic Models, December 2020. URL http://arxiv.org/abs/2006.11239. arXiv:2006.11239 [cs, stat].
  16. [17]Matthew D. Hoffman and Andrew Gelman. The no-U-turn sampler: Adaptively setting path lengths in Hamiltonian Monte Carlo. *Journal of Machine Learning Research*, 15(1):1593–1623, 2014.DOI
  17. [18]Moritz Hoffmann, Martin Konrad Scherer, Tim Hempel, Andreas Mardt, Brian de Silva, Brooke Elena Husic, Stefan Klus, Hao Wu, J Nathan Kutz, Steven Brunton, and Frank Noé. Deeptime: a Python library for machine learning dynamical models from time series data. *Machine Learning: Science and Technology*, 3(1):015009, 2022.DOI
  18. [19]Peter Holderrieth, Douglas Chen, Luca Eyring, Ishin Shah, Giri Anantharaman, Yutong He, Zeynep Akata, Tommi Jaakkola, Nicholas Matthew Boffi, and Max Simchowitz. Diamond Maps: Efficient Reward Alignment via Stochastic Flow Maps, April 2026. URL http://arxiv.org/abs/2602.05993. arXiv:2602.05993 [cs] version: 2.
  19. [20]Scott A. Hollingsworth and Ron O. Dror. Molecular Dynamics Simulation for All. *Neuron*, 99(6):1129–1143, September 2018. doi: 10.1016/j.neuron.2018.08.011. URL https://doi.org/10.1016/j.neuron.2018.08.011. PMID: 30236283; PMCID: PMC6209097.
  20. [21]Koji Hukushima and Koji Nemoto. Exchange monte carlo method and application to spin glass simulations. *Journal of the Physical Society of Japan*, 65(6):1604–1608, 1996. ISSN 1347-4073. doi: 10.1143/jpsj.65.1604. URL http://dx.doi.org/10.1143/JPSJ.65.1604.
  21. [22]John Jumper, Richard Evans, Alexander Pritzel, Tim Green, Michael Figurnov, Olaf Ronneberger, Kathryn Tunyasuvunakool, Russ Bates, Augustin Žídek, Anna Potapenko, Alex Bridgland, Clemens Meyer, Simon A. A. Kohl, Andrew J. Ballard, Andrew Cowie, Bernardino Romera-Paredes, Stanislav Nikolov, Rishub Jain, Jonas Adler, Trevor Back, Stig Petersen, David Reiman, Ellen Clancy, Michal Zielinski, Martin Steinegger, Michalina Pacholska, Tamas Berghammer, Sebastian Bodenstein, David Silver, Oriol Vinyals, Andrew W. Senior, Koray Kavukcuoglu, Pushmeet Kohli, and Demis Hassabis. Highly accurate protein structure prediction with AlphaFold. *Nature*, July 2021. ISSN 0028-0836, 1476-4687. doi: 10.1038/s41586-021-03819-2. URL https://www.nature.com/articles/s41586-021-03819-2.DOI
  22. [23]Sékou-Oumar Kaba, Arnab Kumar Mondal, Yan Zhang, Yoshua Bengio, and Siamak Ravanbakhsh. Equivariance with learned canonicalization functions, 2023. URL https://arxiv.org/abs/2211.06489.
  23. [24]Samuel Karlin and Howard Taylor. *A second course in stochastic processes*. Elsevier, 1981.
  24. [25]Naoki Kiyohara, Edward Johns, and Yingzhen Li. Neural stochastic flows: Solver-free modelling and inference for sde solutions, 2025. URL https://arxiv.org/abs/2510.25769.
  25. [26]Leon Klein and Frank Noé. Transferable Boltzmann Generators, June 2024. URL http://arxiv.org/abs/2406.14426. arXiv:2406.14426 [physics, stat].
  26. [27]Klein, L., Foong, A. Y. K., Fjelde, T. E., Mlodzeniec, B., Brockschmidt, M., Nowozin, S., Noé, F., & Tomioka, R. (2023). Timewarp: Transferable Acceleration of Molecular Dynamics by Learning Time-Coarsened Dynamics. http://arxiv.org/abs/2302.01170
  27. [28]Leon Klein, Andreas Krämer, and Frank Noé. Equivariant flow matching, June 2023b. URL http://arxiv.org/abs/2306.15030. arXiv:2306.15030 [physics, stat].
  28. [29]Kloeden, P. E., & Platen, E. (1992). Numerical solution of stochastic differential equations. <a href="https://doi.org/https://api.semanticscholar.org/CorpusID:116903781" rel="noopener noreferrer" target="_blank">https://doi.org/https://api.semanticscholar.org/CorpusID:116903781</a>
  29. [30]Jonas Köhler, Leon Klein, and Frank Noé. Equivariant flows: Exact likelihood generative learning for symmetric densities. In *International Conference on Machine Learning*, pp. 5361–5370, 2020.arxiv.org/abs/2006.02425
  30. [31]Jonas Köhler, Andreas Krämer, and Frank Noé. Smooth normalizing flows. In *Advances in Neural Information Processing Systems*, volume 34, 2021.arxiv.org/abs/2110.00351
  31. [32]Andrei Kolmogoroff. Über die analytischen methoden in der wahrscheinlichkeitsrechnung. Mathematische Annalen, 104(1):415–458, 1931.DOI
  32. [33]Benedict Leimkuhler and Charles Matthews. Robust and efficient configurational molecular sampling via langevin dynamics. The Journal of chemical physics, 138(17), 2013.DOI
  33. [34]Benedict Leimkuhler and Charles Matthews. Molecular Dynamics: With Deterministic and Stochastic Numerical Methods, volume 39 of Interdisciplinary Applied Mathematics. Springer International Publishing, 2015. ISBN 978-3-319-16375-8. doi: 10.1007/978-3-319-16375-8. URL https://doi.org/10.1007/978-3-319-16375-8.
  34. [36]Yaron Lipman, Ricky T. Q. Chen, Heli Ben-Hamu, Maximilian Nickel, and Matt Le. Flow Matching for Generative Modeling, February 2023. URL http://arxiv.org/abs/2210.02747. arXiv:2210.02747 [cs, stat].
  35. [37]Loshchilov, I., & Hutter, F. (2019). Decoupled weight decay regularization. International Conference on Learning Representations.
  36. [38]Cheng Lu and Yang Song. Simplifying, stabilizing and scaling continuous-time consistency models, 2025. URL https://arxiv.org/abs/2410.11081.
  37. [40]Sam McCallum, Zander W. Blasingame, Timothy Herschell, Niklas Rindtorff, Alexander Tong, and James Foster. Strong stochastic flow maps, 2026. URL https://arxiv.org/abs/2606.01086.
  38. [41]Del Moral, P., & Doucet, A. (2002). Sequential monte carlo samplers. http://arxiv.org/abs/cond-mat/0212648
  39. [42]Frank Noé, Simon Olsson, Jonas Köhler, and Hao Wu. Boltzmann Generators – Sampling Equilibrium States of Many-Body Systems with Deep Learning, July 2019. URL http://arxiv.org/abs/1812.01729. arXiv:1812.01729 [stat].
  40. [43]Alexey Onufriev, Donald Bashford, and David A. Case. Exploring protein native states and large-scale conformational changes with a modified generalized Born model. Proteins: Structure, Function, and Bioinformatics, 55(2):383–394, 2004.pubmed.ncbi.nlm.nih.gov/15048829
  41. [44]Zhengkai Pan, Peter Potaptchik, Wenxi Yao, Michael S. Albergo, and Jakiw Pidstrigach. Itô maps for any-step sdes, 2026. URL https://arxiv.org/abs/2606.11156.
  42. [46]Guillermo Pérez-Hernández, Fabian Paul, Toni Giorgino, Gianni De Fabritiis, and Frank Noé. Identification of slow molecular order parameters for Markov model construction. The Journal of Chemical Physics, 139(1):015102, 2013.DOI
  43. [47]Peter Potaptchik, Cheuk-Kit Lee, and Michael S. Albergo. Tilt matching for scalable sampling and fine-tuning, 2025. URL https://arxiv.org/abs/2512.21829.
  44. [48]Peter Potaptchik, Adhi Saravanan, Abbas Mammadov, Alvaro Prat, Michael S. Albergo, and Yee Whye Teh. Meta Flow Maps enable scalable reward alignment, January 2026. URL http://arxiv.org/abs/2601.14430. arXiv:2601.14430 [stat].
  45. [49]Winfried Ripken, Michael Plainer, Gregor Lied, Thorben Frank, Oliver T. Unke, Stefan Chmiela, Frank Noé, and Klaus-Robert Müller. Learning hamiltonian flow maps: Mean flow consistency large-timestep molecular dynamics, 2026. URL https://arxiv.org/abs/2601.22123.
  46. [50]Gareth O. Roberts and Richard L. Tweedie. Exponential convergence of Langevin distributions and their discrete approximations. *Bernoulli*, 2(4):341–363, 1996.DOI
  47. [51]Mathias Schreiner, Ole Winther, and Simon Olsson. Implicit Transfer Operator Learning: Multiple Time-Resolution Surrogates for Molecular Dynamics, May 2023. URL http://arxiv.org/abs/2305.18046. arXiv:2305.18046 [physics, stat].
  48. [52]Yang Song, Jascha Sohl-Dickstein, Diederik P. Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score-Based Generative Modeling through Stochastic Differential Equations, February 2021. URL http://arxiv.org/abs/2011.13456. arXiv:2011.13456 [cs].
  49. [53]Yang Song, Prafulla Dhariwal, Mark Chen, and Ilya Sutskever. Consistency Models, May 2023. URL http://arxiv.org/abs/2303.01469. arXiv:2303.01469 [cs].
  50. [54]Stroock, D. W., & Varadhan, S. (1979). Multidimensional diffusion processes. <a href="https://doi.org/https://api.semanticscholar.org/CorpusID:118022077" rel="noopener noreferrer" target="_blank">https://doi.org/https://api.semanticscholar.org/CorpusID:118022077</a>
  51. [55]Charlie B. Tan, Avishek Joey Bose, Chen Lin, Leon Klein, Michael M. Bronstein, and Alexander Tong. Scalable Equilibrium Sampling with Sequential Boltzmann Generators, February 2025. URL http://arxiv.org/abs/2502.18462. arXiv:2502.18462 [cs].
  52. [56]Charlie B. Tan, Majdi Hassan, Leon Klein, Saifuddin Syed, Dominique Beaini, Michael M. Bronstein, Alexander Tong, and Kirill Neklyudov. Amortized sampling with transferable normalizing flows, 2026. URL https://arxiv.org/abs/2508.18175.
  53. [57]George Tucker, Dieterich Lawson, Shixiang Gu, and Chris J. Maddison. Doubly reparameterized gradient estimators for Monte Carlo objectives. In *International Conference on Learning Representations*, 2019.arxiv.org/abs/1810.04152
  54. [58]Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2023). Attention Is All You Need. http://arxiv.org/abs/1706.03762
  55. [59]Dongyeop Woo, Marta Skreta, Seonghyun Park, Kirill Neklyudov, and Sungsoo Ahn. Riemannian meanflow. *arXiv preprint arXiv:2602.07744*, 2026.arXiv

Paper details

vlnc:W_k6cjpxbe
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 4

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Computer Science

Not recorded for this paper

View statement
Looking for related papers…
Contents