betaFeedback

WorksPhysics and Astronomy

VESPER: Online Learning of Free Energies with a Multi-Tempered Variational Objective

Abstract

Computing molecular free-energy landscapes over collective variables (CVs) is central to understanding molecular mechanisms, but poses a difficult learning problem without training data available a priori. Samples of the marginal CV distribution must be generated by molecular dynamics (MD), whose trajectories can remain trapped in metastable states and rarely cross the free-energy barriers governing conformational changes or chemical reactions. We introduce VESPER, which couples the generation of these samples to neural-network-driven free-energy learning: MD supplies configurations for learning, while the evolving model biases the MD to accelerate exploration and generate further training data. VESPER minimizes a single variational objective that is convex in the free energy and can combine several tempering levels, each flattening the free-energy landscape to promote barrier crossing. Every level has the same physical free energy as its minimizer, allowing data from all levels to train one shared model without reweighting. Learning uses the difference in mean parameter gradients between biased MD configurations and auxiliary walkers that explore the flattened model in CV space. We prove conditional convergence of the coupled overdamped population dynamics without replica exchange under log-Sobolev and Polyak-Lojasiewicz assumptions. VESPER matches or improves on established enhanced-sampling methods on alanine dipeptide and yields self-consistent free-energy reconstructions from three to six CVs on alanine tetrapeptide. For methane dehydrogenation on Ni(511) with a machine-learned interatomic potential, it recovers two-CV Boltzmann marginals and a reactant-to-saddle free-energy difference along a three-CV path, both qualitatively consistent with the lower-dimensional reference. It also samples RMSD-defined folded and unfolded regions of chignolin while biasing seventeen CVs.

Introduction

Molecular dynamics (MD) is a central tool for studying the structure and dynamics of molecular systems in physics, chemistry, and biology. A recurring goal is to compute the free-energy landscape over a set of informative collective variables (CVs), defined as the negative logarithm of their marginal density under the equilibrium distribution, multiplied by the thermal energy kBTk_{\mathrm{B}}T. This landscape quantifies the relative stability of metastable states and the barriers separating them.

Given samples from this marginal, learning it would be a standard estimation problem. In MD, however, the microscopic potential is known, but the relevant configurations must be generated by simulation. Brute-force MD is often computationally prohibitive: barrier-crossing transitions between metastable states can occur on microsecond-to-millisecond timescales, whereas integration requires femtosecond timesteps [23]. Free-energy learning is therefore coupled to a sampling problem, and the model must accelerate the simulation while it is itself being learned.

We introduce VESPER (Variational Extended Space Parallel Ensemble Regression), which learns a neural free energy GθG_\theta while using it to bias the simulations that produce its own training data. VESPER minimizes a single variational objective: a positively weighted sum of components, one per tempering level, each using a distinct temperature-boosting factor to flatten the effective free-energy landscape. Each component is convex in the free energy and is minimized by the physical free energy at every such level. The sum therefore has the same minimizer up to an additive constant, and every level contributes its gradient to one model without reweighting to the physical ensemble. VESPER couples biased microscopic dynamics, persistent auxiliary CV walkers, and free-energy learning through one shared model. Through the CV map s(x)s(x), the same GθG_\theta sets both the MD bias and the tempered auxiliary potential, with no separate free-energy recursion. We also establish conditional convergence of the coupled sampling and learning dynamics. These contributions build on established variational connections: individual components extend the pseudo-spherical family to negative exponents and admit Rényi variational forms [8, 20, 60], and their gradients recover the instantaneous update of well-tempered variationally enhanced sampling (VES), up to scale, when its target is held fixed during differentiation [9, 58, 59].

The gradient of each component is the difference between two averages of ∇θGθ\nabla_\theta G_\theta: one over MD configurations evolving under the microscopic potential plus a bias built from GθG_\theta, the other over persistent auxiliary walkers evolving directly in CV space under a tempered GθG_\theta. At the optimum both populations sample the same flattened distribution, and the walkers interact with the molecular system only through the shared parameters, so no adiabatic limit is needed to identify the free energy, unlike in temperature-accelerated MD and driven adiabatic free-energy dynamics [1, 40]. Replica exchange can additionally carry configurations discovered under strongly flattened biases down the ladder, facilitating barrier crossing along the CVs.

Our main contributions are:

  • A single objective with a common minimizer. We establish that the losses associated with all tempering levels are convex in the free-energy function and share the physical free energy as their minimizer, up to an additive constant. Their positively weighted sum therefore defines one objective for one model, with every level contributing its native gradient without reweighting to the physical ensemble (Proposition 1).

  • Online learning from coupled sampling populations. Biased MD and persistent auxiliary CV walkers generate the configurations needed for learning as the model evolves, without a pre-existing dataset or an adiabatic limit between microscopic and auxiliary coordinates. We prove conditional convergence of the coupled overdamped population dynamics without replica exchange under log-Sobolev and Polyak–Łojasiewicz assumptions, controlling both objective error and sampling lag (Section A.6).

  • Replica exchange on the learned free energy. The Metropolis acceptance probability for exchange between adjacent levels depends only on learned free-energy differences in lower-dimensional CV space, avoiding additional evaluations of the microscopic potential. At a fixed model, the exchange kernel preserves the product of the sampling laws across levels and therefore leaves the population gradient unchanged.

  • Controlled tests of tempering and replica exchange. On the Müller–Brown potential, the tempering ladder balances basin accuracy, measured by KL divergence, with landscape reconstruction, measured by RMSE. On four-well landscapes spanning two to one hundred dimensions, a fixed ladder remains competitive without dimension-specific selection of α\alpha, with replica exchange improving accuracy and convergence in the most challenging high-dimensional case (§ 3.1).

  • Molecular free-energy reconstruction and high-dimensional exploration. On alanine dipeptide, VESPER achieves competitive reconstruction accuracy and convergence relative to the tested enhanced-sampling baselines. On alanine tetrapeptide, agreement between Boltzmann marginals learned with three and six CVs demonstrates dimensional self-consistency. For methane dehydrogenation on Ni(511) with a machine-learned interatomic potential, VESPER learns a three-CV landscape whose two-CV Boltzmann marginals and reactant-to-saddle free-energy difference along a three-CV path are consistent with the lower-dimensional reference. On chignolin, biasing seventeen dihedral CVs produces repeated RMSD-defined unfolding and refolding transitions, demonstrating effective exploration of a high-dimensional molecular CV space (§ 3).

Methodology

Collective variables and free energy

Let x∈Xx \in\mathcal{X} denote a molecular configuration, U(x)U(x) its microscopic potential energy, and ϱ(x)=Z−1e−βU(x)\varrho(x) = Z^{-1}e^{-\beta U(x)} its equilibrium density at inverse temperature β=(kBT)−1\beta= (k_{\mathrm{B}}T)^{-1}. A collective-variable map s:X→Ss : \mathcal{X} \to\mathcal{S} reduces the dd-dimensional configuration space to a DD-dimensional CV domain, typically with D≪dD \ll d. The equilibrium CV density and its associated free energy are

p(s)=∫Xδ(s−s(x))ϱ(x) dx,G⋆(s)=−β−1log⁡p(s).(1)p(s) = \int_{\mathcal{X}} \delta(s-s(x))\varrho(x)\,\mathrm{d}x,\qquad G_{\star}(s) = -\beta^{-1}\log p(s). \tag*{(1)}

The microscopic potential supplies the forces for simulation, whereas the marginal p(s)p(s) must be inferred from configurations generated by the simulation. VESPER represents G⋆G_{\star} by a differentiable model GθG_{\theta} and uses it to bias the molecular dynamics that generate its training configurations.

The variational objective

Fix an auxiliary inverse temperature 0<β^<β0 < \hat{\beta} < \beta and set α=β^/β∈(0,1)\alpha= \hat{\beta}/\beta\in(0,1). We define the functional

Lβ^[G]=1β^log⁡∫Se−β^G(s) ds+1β−β^log⁡∫Xe(β−β^)G(s(x))ϱ(x) dx.(2)\mathcal{L}_{\hat{\beta}}[G] = \frac{1}{\hat{\beta}}\log\int_{\mathcal{S}}e^{-\hat{\beta}G(s)}\,\mathrm{d}s+\frac{1}{\beta-\hat{\beta}}\log\int_{\mathcal{X}}e^{(\beta-\hat{\beta})G(s(x))}\varrho(x)\,\mathrm{d}x. \tag*{(2)}

The two terms are the log-partition functions of a tempered version of the model on S\mathcal{S} and of the equilibrium density tilted by the model. They define the auxiliary CV law

PG,β^(s)=e−β^G(s)AG,AG=∫Se−β^G(s) ds(3)P_{G,\hat{\beta}}(s)=\frac{e^{-\hat{\beta}G(s)}}{A_G},\qquad A_G=\int_{\mathcal{S}}e^{-\hat{\beta}G(s)}\,\mathrm{d}s \tag*{(3)}

and the biased microscopic law together with its CV marginal,

νG,β^(x)=e(β−β^)G(s(x))ϱ(x)BG,QG,β^(s)=p(s)e(β−β^)G(s)BG,BG=∫Xe(β−β^)G(s(x))ϱ(x) dx.(4)\nu_{G,\hat{\beta}}(x)=\frac{e^{(\beta-\hat{\beta})G(s(x))}\varrho(x)}{B_G},\qquad Q_{G,\hat{\beta}}(s)=\frac{p(s)e^{(\beta-\hat{\beta})G(s)}}{B_G},\qquad B_G=\int_{\mathcal{X}}e^{(\beta-\hat{\beta})G(s(x))}\varrho(x)\,\mathrm{d}x. \tag*{(4)}

The first variation of equation (2) in a direction hh is EQG,β^h−EPG,β^h\mathbb{E}_{Q_{G,\hat{\beta}}}h-\mathbb{E}_{P_{G,\hat{\beta}}}h, so GG is stationary exactly when the two CV laws coincide:

e−β^GAG=pe(β−β^)GBG⟺e−βG∝p⟺G=G⋆+C.(5)\frac{e^{-\hat{\beta}G}}{A_G}=\frac{p e^{(\beta-\hat{\beta})G}}{B_G}\Longleftrightarrow e^{-\beta G}\propto p\Longleftrightarrow G=G_\star+C. \tag*{(5)}

The tempering exponent cancels: for every α\alpha, the stationary point is the physical free energy. The second variation, β^Var⁡PG,β^(h)+(β−β^)Var⁡QG,β^(h)\hat{\beta}\operatorname{Var}_{P_{G,\hat{\beta}}}(h)+(\beta-\hat{\beta})\operatorname{Var}_{Q_{G,\hat{\beta}}}(h), is nonnegative, so Lβ^\mathcal{L}_{\hat{\beta}} is convex and its minimizer set is {G⋆+C:C∈R}\{G_\star+C:C\in\mathbb{R}\} (Proposition S1). At the minimizer,

PG⋆,β^=QG⋆,β^=πα,πα∝e−αβG⋆∝pα,(6)P_{G_\star,\hat{\beta}}=Q_{G_\star,\hat{\beta}}=\pi_\alpha,\qquad\pi_\alpha\propto e^{-\alpha\beta G_\star}\propto p^\alpha, \tag*{(6)}

so both populations explore the same flattened landscape, with barriers reduced by the factor α\alpha. Near the minimizer, Lβ^[G]−Lβ^[G⋆]≈β2Var⁡πα(G−G⋆)\mathcal{L}_{\hat{\beta}}[G]-\mathcal{L}_{\hat{\beta}}[G_\star]\approx\frac{\beta}{2}\operatorname{Var}_{\pi_\alpha}(G-G_\star): α\alpha does not change the target, only where errors are weighted, with smaller α\alpha giving more weight to high free-energy regions.

For a parametric model GθG_\theta, the chain rule gives

∇θLβ^[Gθ]=EνGθ,β^[∇θGθ(s(x))]−EPGθ,β^[∇θGθ(s)].(7)\nabla_\theta\mathcal{L}_{\hat{\beta}}[G_\theta]=\mathbb{E}_{\nu_{G_\theta,\hat{\beta}}}\left[\nabla_\theta G_\theta(s(x))\right]-\mathbb{E}_{P_{G_\theta,\hat{\beta}}}\left[\nabla_\theta G_\theta(s)\right]. \tag*{(7)}

Neither expectation requires knowing pp. The first is over configurations sampled by MD biased with GθG_\theta, and the second over walkers sampling the tempered model in CV space (§ 2.4).

The combined objective

Choose 0≤β^1<⋯<β^K≤β0\leq\hat{\beta}_1<\cdots<\hat{\beta}_K\leq\beta, set αk=β^k/β\alpha_k=\hat{\beta}_k/\beta, and write PG,kP_{G,k}, νG,k\nu_{G,k}, QG,kQ_{G,k} for the laws of level kk. The component Lβ^k\mathcal{L}_{\hat{\beta}_k} is given by equation (2) for 0<β^k<β0<\hat{\beta}_k<\beta. At β^k=β\hat{\beta}_k=\beta and, on compact S\mathcal{S}, at β^k=0\hat{\beta}_k=0, define the endpoint objectives directly under the endpoint conditions in Section A.2:

Lβ[G]=1βlog⁡∫Se−βG(s) ds+Ep[G],L0[G]=−Eu[G]+1βlog⁡Ep[eβG].(8)\mathcal{L}_{\beta}[G] = \frac{1}{\beta} \log\int_{\mathcal{S}} e^{-\beta G(s)} \, \mathrm{d}s + \mathbb{E}_{p}[G], \qquad\mathcal{L}_{0}[G] = -\mathbb{E}_{u}[G] + \frac{1}{\beta} \log\mathbb{E}_{p}[e^{\beta G}]. \tag*{(8)}

References

  1. [1]Jerry B. Abrams and Mark E. Tuckerman. Efficient and direct generation of multidimensional free energy surfaces via adiabatic dynamics without coordinate transformations. The Journal of Physical Chemistry B, 112(49):15742–15757, December 2008. ISSN 1520-6106. doi: 10.1021/jp805039u.DOI
  2. [2]Roman Affentranger, Ivano Tavernelli, and Ernesto E. Di Iorio. A novel Hamiltonian replica exchange MD protocol to enhance protein conformational space sampling. Journal of Chemical Theory and Computation, 2(2):217–228, 2006. doi: 10.1021/ct050250b.DOI
  3. [3]Michael S. Albergo and Eric Vanden-Eijnden. NETS: A non-equilibrium transport sampler. In Proceedings of the 42nd International Conference on Machine Learning, volume 267 of Proceedings of Machine Learning Research, pp. 1026–1055. PMLR, 2025. URL https://proceedings.mlr.press/v267/albergo25a.html.arxiv.org/abs/2410.02711
  4. [4]Shalini Awasthi and Nisanth N. Nair. Exploring high dimensional free energy landscapes: Temperature accelerated sliced sampling. The Journal of Chemical Physics, 146(9):094108, 2017. doi: 10.1063/1.4977704.DOI
  5. [5]Shitanshu Bajpai, Brian K. Petkov, Muchen Tong, Charles R. A. Abreu, Nisanth N. Nair, and Mark E. Tuckerman. An interoperable implementation of collective-variable based enhanced sampling methods in extended phase space within the OpenMM package. Journal of Computational Chemistry, 44(28):2166–2183, 2023. ISSN 1096-987X. doi: 10.1002/jcc.27182.DOI
  6. [6]Alessandro Barducci, Giovanni Bussi, and Michele Parrinello. Well-tempered metadynamics: A smoothly converging and tunable free-energy method. Physical Review Letters, 100(2):020603, January 2008. doi: 10.1103/PhysRevLett.100.020603.DOI
  7. [7]Simon Batzner, Albert Musaelian, Lixin Sun, Mario Geiger, Jonathan P. Mailoa, Mordechai Kornbluth, Nicola Molinari, Tess E. Smidt, and Boris Kozinsky. E(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials. Nature Communications, 13(1):2453, 2022. doi: 10.1038/s41467-022-29939-5.DOI
  8. [8]Jeremiah Birrell, Paul Dupuis, Markos A. Katsoulakis, Luc Rey-Bellet, and Jie Wang. Variational representations and neural network estimation of Rényi divergences. SIAM Journal on Mathematics of Data Science, 3(4):1093–1116, 2021. doi: 10.1137/20M1368926.DOI
  9. [9]Luigi Bonati, Yue-Yu Zhang, and Michele Parrinello. Neural networks-based variationally enhanced sampling. Proceedings of the National Academy of Sciences, 116(36):17641–17647, September 2019. doi: 10.1073/pnas.1907975116.DOI
  10. [10]Davide Carbone, Mengjian Hua, Simon Coste, and Eric Vanden-Eijnden. Efficient training of energy-based models using Jarzynski equality. In Advances in Neural Information Processing Systems, volume 36, pp. 52583–52614, 2023. doi: 10.52202/075280-2290. URL https://proceedings.neurips.cc/paper_files/paper/2023/hash/a4ddb865e0a8ca3cca43fd7387b4b0da-Abstract-Conference.html.DOI
  11. [11]Ming Chen, Michel A. Cuendet, and Mark E. Tuckerman. Heating and flooding: a unified approach for rapid generation of free energy surfaces. The Journal of Chemical Physics, 137(2):024102, 2012. doi: 10.1063/1.4733389.DOI
  12. [12]KyungHyun Cho, Tapani Raiko, and Alexander Ilin. Parallel tempering is efficient for learning restricted Boltzmann machines. In Proceedings of the International Joint Conference on Neural Networks (IJCNN), pp. 1–8, 2010. doi: 10.1109/IJCNN.2010.5596837.DOI
  13. [13]Eric Darve and Andrew Pohorille. Calculating free energies using average force. The Journal of Chemical Physics, 115(20):9169–9183, November 2001. ISSN 0021-9606. doi: 10.1063/1.1410978.DOI
  14. [14]Eric Darve, David Rodríguez-Gómez, and Andrew Pohorille. Adaptive biasing force method for scalar and vector free energy calculations. The Journal of Chemical Physics, 128(14):144120, April 2008. ISSN 0021-9606. doi: 10.1063/1.2829861.DOI
  15. [15]Guillaume Desjardins, Aaron Courville, Yoshua Bengio, Pascal Vincent, and Olivier Dellalleau. Tempered Markov chain Monte Carlo for training of restricted Boltzmann machines. In Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics (AISTATS), volume 9 of Proceedings of Machine Learning Research, pp. 145–152, 2010. URL https://proceedings.mlr.press/v9/desjardins10a.html.
  16. [16]Manuel Dibak, Leon Klein, Andreas Krämer, and Frank Noé. Temperature steerable flows and Boltzmann generators. Physical Review Research, 4(4):L042005, 2022. doi: 10.1103/PhysRevResearch.4.L042005.DOI
  17. [17]Weinan E, Weiqing Ren, and Eric Vanden-Eijnden. Simplified and improved string method for computing the minimum energy paths in barrier-crossing events. The Journal of Chemical Physics, 126(16):164103, April 2007. ISSN 0021-9606. doi: 10.1063/1.2720838.
  18. [18]David J. Earl and Michael W. Deem. Parallel tempering: Theory, applications, and new perspectives. Physical Chemistry Chemical Physics, 7(23):3910–3916, 2005. doi: 10.1039/B509983H.DOI
  19. [19]Peter Eastman, Jason Swails, John D. Chodera, Robert T. McGibbon, Yutong Zhao, Kyle A. Beauchamp, Lee-Ping Wang, Andrew C. Simmonett, Matthew P. Harrigan, Chaya D. Stern, Rafal P. Wiewiora, Bernard R. Brooks, and Vijay S. Pande. OpenMM 7: Rapid development of high performance algorithms for molecular dynamics. PLOS Computational Biology, 13(7):e1005659, 2017. doi: 10.1371/journal.pcbi.1005659.DOI
  20. [20]Hironori Fujisawa and Shinto Eguchi. Robust parameter estimation with a small bias against heavy contamination. Journal of Multivariate Analysis, 99(9):2053–2081, 2008. doi: 10.1016/j.jmva.2008.02.004.DOI
  21. [21]Alejandro Gil-Ley and Giovanni Bussi. Enhanced conformational sampling using replica exchange with collective-variable tempering. Journal of Chemical Theory and Computation, 11 (3):1077–1085, 2015. doi: 10.1021/ct5009087. Correction: J. Chem. Theory Comput. 11, 5554 (2015), doi:10.1021/acs.jctc.5b00981.
  22. [22]Ashley Z. Guo, Emre Sevgen, Hythem Sidky, Jonathan K. Whitmer, Jeffrey A. Hubbell, and Juan J. de Pablo. Adaptive enhanced sampling by force-biasing using neural networks. The Journal of Chemical Physics, 148(13):134108, 2018. doi: 10.1063/1.5020733.DOI
  23. [23]Jérôme Hénin, Tony Lelièvre, Michael R. Shirts, Omar Valsson, and Lucie Delemotte. Enhanced sampling methods for molecular dynamics simulations. Living Journal of Computational Molecular Science, 4(1):1583, 2022. doi: 10.33011/livecoms.4.1.1583.DOI
  24. [24]Geoffrey E. Hinton. Training products of experts by minimizing contrastive divergence. Neural Computation, 14(8):1771–1800, 2002. doi: 10.1162/089976602760276018.pubmed.ncbi.nlm.nih.gov/12180402
  25. [25]Richard Holley and Daniel Stroock. Logarithmic Sobolev inequalities and stochastic Ising models. Journal of Statistical Physics, 46(5-6):1159–1194, 1987. doi: 10.1007/BF01011161.DOI
  26. [26]Shinya Honda, Toshihiko Akiba, Yusuke S. Kato, Yoshito Sawada, Masakazu Sekijima, Miyuki Ishimura, Ayako Ooishi, Hideki Watanabe, Takayuki Odahara, and Kazuaki Harata. Crystal structure of a ten-amino acid protein. Journal of the American Chemical Society, 130(46):15327–15331, 2008. doi: 10.1021/ja8030533.DOI
  27. [27]Viktor Hornak, Robert Abel, Asim Okur, Bentley Strockbine, Adrian Roitberg, and Carlos Simmerling. Comparison of multiple Amber force fields and development of improved protein backbone parameters. Proteins: Structure, Function, and Bioinformatics, 65(3):712–725, 2006. doi: 10.1002/prot.21123.DOI
  28. [28]William Humphrey, Andrew Dalke, and Klaus Schulten. VMD: Visual molecular dynamics. Journal of Molecular Graphics, 14(1):33–38, 1996. doi: 10.1016/0263-7855(96)00018-5.DOI
  29. [29]Michele Invernizzi and Michele Parrinello. Rethinking metadynamics: From bias potentials to probability distributions. The Journal of Physical Chemistry Letters, 11(7):2731–2736, 2020. doi: 10.1021/acs.jpclett.0c00497.DOI
  30. [30]Michele Invernizzi, Andreas Krämer, Cecilia Clementi, and Frank Noé. Skipping the replica exchange ladder with normalizing flows. The Journal of Physical Chemistry Letters, 13(50):11643–11649, 2022. doi: 10.1021/acs.jpclett.2c03327.DOI
  31. [31]Pavel Izmailov, Dmitrii Podoprikhin, Timur Garipov, Dmitry Vetrov, and Andrew Gordon Wilson. Averaging weights leads to wider optima and better generalization. In Proceedings of the 34th Conference on Uncertainty in Artificial Intelligence (UAI), volume R16 of Proceedings of Machine Learning Research, pp. 875–884. PMLR, 2018. URL https://proceedings.mlr.press/r16/izmailov18a.html.arxiv.org/abs/1803.05407
  32. [32]Christopher Jarzynski. Nonequilibrium equality for free energy differences. Physical Review Letters, 78(14):2690–2693, 1997. doi: 10.1103/PhysRevLett.78.2690.DOI
  33. [33]Yezhi Jin, Yinan Xu, Jireh S. García Sánchez, Gustavo R. Pérez-Lemus, Pablo F. Zubieta Rico, Massimiliano Delferro, and Juan J. de Pablo. Free-energy landscapes and surface dynamics in methane activation on Ni(511) via machine learning and enhanced sampling. ACS Catalysis, 15(11):8931–8942, June 2025. doi: 10.1021/acscatal.5c00724.DOI
  34. [34]William L. Jorgensen, Jayaraman Chandrasekhar, Jeffry D. Madura, Roger W. Impey, and Michael L. Klein. Comparison of simple potential functions for simulating liquid water. The Journal of Chemical Physics, 79(2):926–935, 1983. doi: 10.1063/1.445869.DOI
  35. [35]Wolfgang Kabsch. A solution for the best rotation to relate two sets of vectors. Acta Crystallographica Section A, 32(5):922–923, 1976. doi: 10.1107/S0567739476001873.DOI
  36. [36]Alessandro Laio and Michele Parrinello. Escaping free-energy minima. Proceedings of the National Academy of Sciences, 99(20):12562–12566, October 2002. doi: 10.1073/pnas.202427399.DOI
  37. [37]Kaizhao Liang, Lizhang Chen, Bo Liu, and Qiang Liu. Cautious optimizers: Improving training with one line of code. In International Conference on Learning Representations, 2026. URL https://proceedings.iclr.cc/paper_files/paper/2026/hash/ae03bdef276132fae089692445725635-Abstract-Conference.html.
  38. [38]Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. In International Conference on Learning Representations, 2019. URL https://openreview.net/forum?id=Bkg6RiCqY7.
  39. [39]James A. Maier, Carmenza Martinez, Koushik Kasavajhala, Lauren Wickstrom, Kevin E. Hauser, and Carlos Simmerling. ff14SB: Improving the accuracy of protein side chain and backbone parameters from ff99SB. Journal of Chemical Theory and Computation, 11(8): 3696–3713, 2015. doi: 10.1021/acs.jctc.5b00255.DOI
  40. [40]Luca Maragliano and Eric Vanden-Eijnden. A temperature accelerated method for sampling free energy and determining reaction pathways in rare events simulations. Chemical Physics Letters, 426(1-3):168–175, July 2006. ISSN 0009-2614. doi: 10.1016/j.cplett.2006.05.062.DOI
  41. [41]Luca Maragliano and Eric Vanden-Eijnden. Single-sweep methods for free energy calculations. The Journal of Chemical Physics, 128(18):184110, 2008. doi: 10.1063/1.2907241.DOI
  42. [42]Klaus Müller and Leo D. Brown. Location of saddle points and minimum energy paths by a constrained simplex optimization procedure. Theoretica Chimica Acta, 53(1):75–93, 1979. doi: 10.1007/BF00547608.
  43. [43]Albert Musaelian, Simon Batzner, Anders Johansson, Lixin Sun, Cameron J. Owen, Mordechai Kornbluth, and Boris Kozinsky. Learning local equivariant representations for large-scale atomistic dynamics. Nature Communications, 14(1):579, 2023. doi: 10.1038/s41467-023-36329-y.DOI
  44. [44]Radford M. Neal. Annealed importance sampling. Statistics and Computing, 11(2):125–139, 2001. doi: 10.1023/A:1008923215028.DOI
  45. [45]Frank Noé, Simon Olsson, Jonas Köhler, and Hao Wu. Boltzmann generators: Sampling equilibrium states of many-body systems with deep learning. Science, 365(6457):eaaw1147, 2019. doi: 10.1126/science.aaw1147.
  46. [46]Stefano Piana and Alessandro Laio. A bias-exchange approach to protein folding. The Journal of Physical Chemistry B, 111(17):4553–4559, 2007. doi: 10.1021/jp067873l.DOI
  47. [47]Boris T. Polyak and Anatoli B. Juditsky. Acceleration of stochastic approximation by averaging. SIAM Journal on Control and Optimization, 30(4):838–855, 1992. doi: 10.1137/0330046.DOI
  48. [48]Lula Rosso, Peter Mináry, Zhongwei Zhu, and Mark E. Tuckerman. On the use of the adiabatic molecular dynamics technique in the calculation of free energy profiles. The Journal of Chemical Physics, 116(11):4389–4402, 2002. doi: 10.1063/1.1448491.DOI
  49. [49]David Ruppert. Efficient estimations from a slowly convergent Robbins-Monro process. Technical Report 781, School of Operations Research and Industrial Engineering, Cornell University, 1988. URL https://hdl.handle.net/1813/8664.
  50. [50]Elia Schneider, Luke Dai, Robert Q. Topper, Christof Drechsel-Grau, and Mark E. Tuckerman. Stochastic neural network approach for learning high-dimensional free energy surfaces. Physical Review Letters, 119(15):150601, 2017. doi: 10.1103/PhysRevLett.119.150601.DOI
  51. [51]Emre Sevgen, Ashley Z. Guo, Hythem Sidky, Jonathan K. Whitmer, and Juan J. de Pablo. Combined force-frequency sampling for simulation of systems having rugged free energy landscapes. Journal of Chemical Theory and Computation, 16(3):1448–1455, 2020. doi: 10.1021/acs.jctc.9b00883.DOI
  52. [52]Patrick Shaffer, Omar Valsson, and Michele Parrinello. Enhanced, targeted sampling of high-dimensional free-energy landscapes using variationally enhanced sampling, with an application to chignolin. Proceedings of the National Academy of Sciences, 113(5):1150–1155, 2016. doi: 10.1073/pnas.1519712113.DOI
  53. [53]Hythem Sidky and Jonathan K. Whitmer. Learning free energy landscapes using artificial neural networks. The Journal of Chemical Physics, 148(10):104111, 2018. doi: 10.1063/1.5018708.DOI
  54. [54]Vincent Sitzmann, Julien N. P. Martel, Alexander W. Bergman, David B. Lindell, and Gordon Wetzstein. Implicit neural representations with periodic activation functions. In Advances in Neural Information Processing Systems, volume 33, pp. 7462–7473, 2020. URL https://proceedings.neurips.cc/paper_files/paper/2020/hash/53c04118df112c13a8c34b38343b9c10-Abstract.html.arxiv.org/abs/2006.09661
  55. [55]Yuji Sugita and Yuko Okamoto. Replica-exchange molecular dynamics method for protein folding. Chemical Physics Letters, 314(1-2):141–151, 1999. ISSN 0009-2614. doi: 10.1016/S0009-2614(99)01123-9.DOI
  56. [56]Tijmen Tieleman. Training restricted Boltzmann machines using approximations to the likelihood gradient. In Proceedings of the 25th International Conference on Machine Learning (ICML), pp. 1064–1071, 2008. doi: 10.1145/1390156.1390290.DOI
  57. [57]Glenn M. Torrie and John P. Valleau. Nonphysical sampling distributions in Monte Carlo free-energy estimation: Umbrella sampling. Journal of Computational Physics, 23(2):187–199, February 1977. ISSN 0021-9991. doi: 10.1016/0021-9991(77)90121-8.DOI
  58. [58]Omar Valsson and Michele Parrinello. Variational approach to enhanced sampling and free energy calculations. Physical Review Letters, 113(9):090601, August 2014. ISSN 1079-7114. doi: 10.1103/PhysRevLett.113.090601.DOI
  59. [59]Omar Valsson and Michele Parrinello. Well-tempered variational approach to enhanced sampling. Journal of Chemical Theory and Computation, 11(5):1996–2002, 2015. doi: 10.1021/acs.jctc.5b00076.DOI
  60. [60]Lantao Yu, Jiaming Song, Yang Song, and Stefano Ermon. Pseudo-spherical contrastive divergence. In Advances in Neural Information Processing Systems, volume 34, pp. 22348–22362, 2021. URL https://proceedings.neurips.cc/paper_files/paper/2021/file/bc5fcb0018cecacba559dc512740091b-Paper.pdf.arxiv.org/abs/2111.00780
  61. [61]Linfeng Zhang, Han Wang, and Weinan E. Reinforced dynamics for enhanced sampling in large atomic and molecular systems. The Journal of Chemical Physics, 148(12):124113, 2018. doi: 10.1063/1.5019675.DOI
  62. [62]Pablo F. Zubieta Rico, Ludwig Schneider, Gustavo R. Pérez-Lemus, Riccardo Alessandri, Siva Dasetty, Trung D. Nguyen, Cintia A. Menéndez, Yiheng Wu, Yezhi Jin, Yinan Xu, Samuel Varner, John A. Parker, Andrew L. Ferguson, Jonathan K. Whitmer, and Juan J. de Pablo. PyVSAGES: flexible, advanced sampling methods accelerated with GPUs. npj Computational Materials, 10(1):35, February 2024. ISSN 2057-3960. doi: 10.1038/s41524-023-01189-z.DOI
  63. [63]Pablo F. Zubieta Rico, Gustavo R. Pérez-Lemus, and Juan J. de Pablo. Efficient sampling of free energy landscapes with functions in Sobolev spaces. The Journal of Chemical Physics, 162 (8):084109, 2025. doi: 10.1063/5.0221263.DOI

Paper details

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

views · 0

downloads · 0

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

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents