Differentiable Molecular Simulations for Control and Learning

Abstract

Molecular dynamics simulations use statistical mechanics at the atomistic scale to enable both the elucidation of fundamental mechanisms and the engineering of matter for desired tasks. The behavior of molecular systems at the microscale is typically simulated with differential equations parameterized by a Hamiltonian, or energy function. The Hamiltonian describes the state of the system and its interactions with the environment. In order to derive predictive microscopic models, one wishes to infer a molecular Hamiltonian that agrees with observed macroscopic quantities. From the perspective of engineering, one wishes to control the Hamiltonian to achieve desired simulation outcomes and structures, as in self-assembly and optical control, to then realize systems with the desired Hamiltonian in the lab. In both cases, the goal is to modify the Hamiltonian such that emergent properties of the simulated system match a given target. We demonstrate how this can be achieved using differentiable simulations where bulk target observables and simulation outcomes can be analytically differentiated with respect to Hamiltonians, opening up new routes for parameterizing Hamiltonians to infer macroscopic models and develop control protocols.

arXiv comment14 pages, 6 figures

References in corpus (52)

• TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems
• A nonequilibrium equality for free energy differences
• Neural Message Passing for Quantum Chemistry
• Gaussian Approximation Potentials: the accuracy of quantum mechanics, without the electrons
• The Entropy Production Fluctuation Theorem and the Nonequilibrium Work Relation for Free Energy Differences
• DeepXDE: A deep learning library for solving differential equations
• Deep Potential Molecular Dynamics: a scalable model with the accuracy of quantum mechanics
• ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost
• Solving high-dimensional partial differential equations using deep learning
• Quantum-Chemical Insights from Deep Tensor Neural Networks
• Equilibrium free energy differences from nonequilibrium measurements: a master equation approach
• Interaction Networks for Learning about Objects, Relations and Physics
• PDE-Net 2.0: Learning PDEs from Data with A Numeric-Symbolic Hybrid Deep Network
• Learning data driven discretizations for partial differential equations
• Time-lagged autoencoders: Deep learning of slow collective variables for molecular kinetics
• Differentiable Programming Tensor Networks
• Variational Encoding of Complex Dynamics
• Unsupervised machine learning in atomistic simulations, between predictions and understanding
• Neural networks-based variationally enhanced sampling
• Graph Dynamical Networks for Unsupervised Learning of Atomic Scale Dynamics in Materials
• Symplectic ODE-Net: Learning Hamiltonian Dynamics with Control
• Rational design of self-assembly pathways for complex multicomponent structures
• Escorted Free Energy Simulations: Improving Convergence by Reducing Dissipation
• Deep Dynamical Modeling and Control of Unsteady Fluid Flows
• Targeted free energy perturbation
• Neural network based path collective variables for enhanced sampling of phase transformations
• Design principles for non-equilibrium self-assembly
• A geometric approach to optimal nonequilibrium control: Minimizing dissipation in nanomagnetic spin systems
• Hamiltonian Graph Networks with ODE Integrators
• DiffTaichi: Differentiable Programming for Physical Simulation
• Boltzmann Generators -- Sampling Equilibrium States of Many-Body Systems with Deep Learning
• Thermodynamic geometry of minimum-dissipation driven barrier crossing
• Principal component analysis of nonequilibrium molecular dynamics simulations
• Adversarial-Residual-Coarse-Graining: Applying machine learning theory to systematic molecular coarse-graining
• Learning Compositional Koopman Operators for Model-Based Control
• Optimal thermodynamic control in open quantum systems
• Theoretical prediction of free-energy landscapes for complex self-assembly
• End-to-End Learning on 3D Protein Structure for Interface Prediction
• Monge-Ampère Flow for Generative Modeling
• SPNets: Differentiable Fluid Dynamics for Deep Neural Networks
• Direct observation and rational design of nucleation behavior in addressable self-assembly
• Sandwiching the marginal likelihood using bidirectional Monte Carlo
• Neural Canonical Transformation with Symplectic Flows
• Controlled exploration of chemical space by machine learning of coarse-grained representations
• Large deviations and optimal control forces for hard particles in one dimension
• Variational approach to the optimal control of coherently driven, open quantum system dynamics
• Model Reduction with Memory and the Machine Learning of Dynamical Systems
• Dynamical computation of the density of states and Bayes factors using nonequilibrium importance sampling
• An efficient approach to the quantum dynamics and rates of processes induced by natural incoherent light
• Machine Learning of coarse-grained Molecular Dynamics Force Fields
• Ultrashort pulse two-photon coherent control of a macroscopic phenomena: light-induced current from channelrhodopsin-2 in live brain cells
• Considerations regarding one-photon phase control

Cited by in corpus (12)

• Solver-in-the-Loop: Learning from Differentiable Physics to Interact with Iterative PDE-Solvers
• Learning neural network potentials from experimental data via Differentiable Trajectory Reweighting
• End-to-End Differentiable Molecular Mechanics Force Field Construction
• Graph Neural Networks Accelerated Molecular Dynamics
• Temperature-transferable coarse-graining of ionic liquids with dual graph convolutional neural networks
• Learning Differential Equations that are Easy to Solve
• Learning Pair Potentials using Differentiable Simulations
• On the design space between molecular mechanics and machine learning force fields
• Enhanced sampling of robust molecular datasets with uncertainty-based collective variables
• A composable autoencoder-based iterative algorithm for accelerating numerical simulations
• A Latent space solver for PDE generalization
• Algorithmic differentiation for plane-wave DFT: materials design, error control and learning model parameters