-
Latent Thermodynamic Flows: Unified Representation Learning and Generative Modeling of Temperature-Dependent Behaviors from Limited Data
Authors:
Yunrui Qiu,
Richard John,
Lukas Herron,
Pratyush Tiwary
Abstract:
Accurate characterization of the equilibrium distributions of complex molecular systems and their dependence on environmental factors such as temperature is essential for understanding thermodynamic properties and transition mechanisms. Projecting these distributions onto meaningful low-dimensional representations enables interpretability and downstream analysis. Recent advances in generative AI,…
▽ More
Accurate characterization of the equilibrium distributions of complex molecular systems and their dependence on environmental factors such as temperature is essential for understanding thermodynamic properties and transition mechanisms. Projecting these distributions onto meaningful low-dimensional representations enables interpretability and downstream analysis. Recent advances in generative AI, particularly flow models such as Normalizing Flows (NFs), have shown promise in modeling such distributions, but their scope is limited without tailored representation learning. In this work, we introduce Latent Thermodynamic Flows (LaTF), an end-to-end framework that tightly integrates representation learning and generative modeling. LaTF unifies the State Predictive Information Bottleneck (SPIB) with NFs to simultaneously learn low-dimensional latent representations, referred to as Collective Variables (CVs), classify metastable states, and generate equilibrium distributions across temperatures beyond the training data. The two components of representation learning and generative modeling are optimized jointly, ensuring that the learned latent features capture the system's slow, important degrees of freedom while the generative model accurately reproduces the system's equilibrium behavior. We demonstrate LaTF's effectiveness across diverse systems, including a model potential, the Chignolin protein, and cluster of Lennard Jones particles, with thorough evaluations and benchmarking using multiple metrics and extensive simulations. Finally, we apply LaTF to a RNA tetraloop system, where despite using simulation data from only two temperatures, LaTF reconstructs the temperature-dependent structural ensemble and melting behavior, consistent with experimental and prior extensive computational results.
△ Less
Submitted 3 July, 2025;
originally announced July 2025.
-
A survey of probabilistic generative frameworks for molecular simulations
Authors:
Richard John,
Lukas Herron,
Pratyush Tiwary
Abstract:
Generative artificial intelligence is now a widely used tool in molecular science. Despite the popularity of probabilistic generative models, numerical experiments benchmarking their performance on molecular data are lacking. In this work, we introduce and explain several classes of generative models, broadly sorted into two categories: flow-based models and diffusion models. We select three repre…
▽ More
Generative artificial intelligence is now a widely used tool in molecular science. Despite the popularity of probabilistic generative models, numerical experiments benchmarking their performance on molecular data are lacking. In this work, we introduce and explain several classes of generative models, broadly sorted into two categories: flow-based models and diffusion models. We select three representative models: Neural Spline Flows, Conditional Flow Matching, and Denoising Diffusion Probabilistic Models, and examine their accuracy, computational cost, and generation speed across datasets with tunable dimensionality, complexity, and modal asymmetry. Our findings are varied, with no one framework being the best for all purposes. In a nutshell, (i) Neural Spline Flows do best at capturing mode asymmetry present in low-dimensional data, (ii) Conditional Flow Matching outperforms other models for high-dimensional data with low complexity, and (iii) Denoising Diffusion Probabilistic Models appears the best for low-dimensional data with high complexity. Our datasets include a Gaussian mixture model and the dihedral torsion angle distribution of the Aib\textsubscript{9} peptide, generated via a molecular dynamics simulation. We hope our taxonomy of probabilistic generative frameworks and numerical results may guide model selection for a wide range of molecular tasks.
△ Less
Submitted 14 November, 2024;
originally announced November 2024.
-
Generative artificial intelligence for computational chemistry: a roadmap to predicting emergent phenomena
Authors:
Pratyush Tiwary,
Lukas Herron,
Richard John,
Suemin Lee,
Disha Sanwal,
Ruiyu Wang
Abstract:
The recent surge in Generative Artificial Intelligence (AI) has introduced exciting possibilities for computational chemistry. Generative AI methods have made significant progress in sampling molecular structures across chemical species, developing force fields, and speeding up simulations. This Perspective offers a structured overview, beginning with the fundamental theoretical concepts in both G…
▽ More
The recent surge in Generative Artificial Intelligence (AI) has introduced exciting possibilities for computational chemistry. Generative AI methods have made significant progress in sampling molecular structures across chemical species, developing force fields, and speeding up simulations. This Perspective offers a structured overview, beginning with the fundamental theoretical concepts in both Generative AI and computational chemistry. It then covers widely used Generative AI methods, including autoencoders, generative adversarial networks, reinforcement learning, flow models and language models, and highlights their selected applications in diverse areas including force field development, and protein/RNA structure prediction. A key focus is on the challenges these methods face before they become truly predictive, particularly in predicting emergent chemical phenomena. We believe that the ultimate goal of a simulation method or theory is to predict phenomena not seen before, and that Generative AI should be subject to these same standards before it is deemed useful for chemistry. We suggest that to overcome these challenges, future AI models need to integrate core chemical principles, especially from statistical mechanics.
△ Less
Submitted 4 September, 2024;
originally announced September 2024.
-
Magnetization switching of FePt nanoparticle recording medium by femtosecond laser pulses
Authors:
R. John,
M. Berritta,
D. Hinzke,
C. Müller,
T. Santos,
H. Ulrichs,
P. Nieves,
J. Walowski,
R. Mondal,
O. Chubykalo-Fesenko,
J. McCord,
P. M. Oppeneer,
U. Nowak,
M. Münzenberg
Abstract:
Manipulation of magnetization with ultrashort laser pulses is promising for information storage device applications. The dynamic of the magnetization response depends on the energy transfer from the photons to the spins during the initial laser excitation. A material of special interest for magnetic storage is FePt nanoparticles , on which optical writing with optical angular momentum was demonstr…
▽ More
Manipulation of magnetization with ultrashort laser pulses is promising for information storage device applications. The dynamic of the magnetization response depends on the energy transfer from the photons to the spins during the initial laser excitation. A material of special interest for magnetic storage is FePt nanoparticles , on which optical writing with optical angular momentum was demonstrated recently by Lambert et al., although the mechanism remained unclear. Here we investigate experimentally and theoretically the all-optical switching of FePt nanoparticles. We show that the magnetization switching is a stochastic process. We develop a complete multiscale model which allows us to optimize the number of laser shots needed to write the magnetization of high anisotropy FePt nanoparticles in our experiments. We conclude that only angular momentum induced optically by the inverse Faraday effect will provide switching with one single femtosecond laser pulse.
△ Less
Submitted 28 June, 2016;
originally announced June 2016.
-
Single- and few-layer graphene growth on stainless steel substrates by direct thermal chemical vapor deposition
Authors:
Robin John,
A Ashokreddy,
C Vijayan,
T Pradeep
Abstract:
Steeping interest on graphene research in basic sciences and applications emphasizes the need for an economical means of synthesizing it. We report a method for the synthesis of graphene on commercially available stainless steel foils using direct thermal chemical vapor deposition. Our method of synthesis and the use of relatively cheap precursors such as ethanol (CH3CH2OH) as a source of carbon a…
▽ More
Steeping interest on graphene research in basic sciences and applications emphasizes the need for an economical means of synthesizing it. We report a method for the synthesis of graphene on commercially available stainless steel foils using direct thermal chemical vapor deposition. Our method of synthesis and the use of relatively cheap precursors such as ethanol (CH3CH2OH) as a source of carbon and SS 304 as the substrate, proved to be economically viable. Presence of single- and few-layer graphene was confirmed using confocal Raman microscopy/spectroscopy. X-ray photoelectron spectroscopic measurements were further used to establish the influence of various elemental species present in stainless steel on graphene growth. Role of cooling rate on surface migration of certain chemical species (oxides of Fe, Cr and Mn) that promote or hinder the growth of graphene is probed. Such analysis of the chemical species present on the surface can be promising for graphene based catalytic research.
△ Less
Submitted 21 September, 2011; v1 submitted 13 August, 2010;
originally announced August 2010.
