Simulating the Interface between Chemistry and Physics,
One Atom at a Time.
Path Integral Molecular Dynamics
​We develop classical molecular dynamics (MD) simulations that provide quantum expectation values for condensed phase systems.
​
Enhanced Sampling
We develop methods to investigate slow chemical transformations, such as protein folding or crystal nucleation and growth using molecular simulations.
​
Machine Learning and Molecular Simulations
We apply powerful machine learning (ML) tools, such as artificial neural networks, to push the boundaries of molecular simulations.
Latest Publications
Adaptive resetting for informed search strategies and the design of non-equilibrium steady-states
Keidar, T.D.; Blumer, O.; Hirshberg, B.; Reuveni, S.
Nature Communications
2025
Periodic boundary conditions for bosonic path integral molecular dynamics
Higer, J.; Feldman, Y.M.Y.; Hirshberg, B.
Journal of Chemical Physics
2025
First-passage approach to optimizing perturbations for improved training of machine learning models
Meir S.; Keidar T.D.; Reuveni S.; Hirshberg B.
Machine Learning: Science and Technology
2025
Have You Tried Turning It Off and On Again? Stochastic Resetting for Enhanced Sampling
Blumer, O.; Hirshberg, B.
Wiley Interdisciplinary Reviews: Computational Molecular Science
2025
Research spotlight:
Path integral molecular dynamics including exchange effects.
Being bosons or fermions is one of the most fundamental properties of quantum particles. PIMD simulations allow us to study thermodynamic properties of quantum systems using classical molecular dynamics - a great computational advantage. However, they completely neglect the effects of exchange. We have recently solved this important problem by developing a new PIMD method for bosons and fermions.
​
Research spotlight:
Stochastic Resetting for Enhanced Sampling
Molecular dynamics simulations provide insights into various processes and have different technological applications, including drug design. However, these simulations are limited to processes slower than one-millionth of a second, and cannot describe slower processes such as protein folding and crystal nucleation. Counter-intuitively, we showed that we can accelerate simulations through randomly stopping, and restarting them.
News
We are hiring!
Several positions are available at all levels (Postdoc, Ph.D., MSc, undergraduate interns).
If you are interested in discovering how machine learning can improve molecular simulations, how to study quantum condensed phase systems using classical simulations, or how to apply these tools to exotic quantum materials or conformational sampling in biomolecules, get in touch!
To apply, please send your CV and a short description of your research interests to Barak.