2024, JCTC, AdaptiveBandit: A Multi-armed Bandit Framework for Adaptive Sampling in Molecular Simulations
Paper Information
Title. AdaptiveBandit: A Multi-armed Bandit Framework for Adaptive Sampling in Molecular Simulations
Year. 2020-06-15
DOI Link. https://doi.org/10.1021/acs.jctc.0c00205
Author(s). Gianni De Fabritiis (Computational Science Laboratory, U Pompeu Fabra, Barcelona, Spain), Adria Perez (Computational Science Laboratory, U Pompeu Fabra, Barcelona, Spain), Pablo Herrera-Nieto (Computational Science Laboratory, U Pompeu Fabra, Barcelona, Spain), Stefan Doerr (Computational Science Laboratory, U Pompeu Fabra, Barcelona, Spain)
Journal. JCTC (Journal of Chemical Theory and Computations)
Volume. 16
Issue. 7
Pages. 4685-4693
CiteKey. Perez2020
ItemType. Implementation
[!https://pubs.acs.org/cms/10.1021/acs.jctc.0c00205/asset/images/medium/ct0c00205_0005.gif]
Abstract
- Sampling from the Equilibrium Distribution has always been a major problem in Molecular Simulations due to the very high Dimensionality of the Conformational Space.
- Over several decades, many approaches have been used to overcome this problem
- In particular, we focus on Unbiased Simulation Methods such as Parallel and Adaptive Sampling
- Here, we recast Adaptive Sampling Scgenes on the basis of Multi-armed Bandits and Develop a Novel Adaptive Sampling Algorithm under this Framework, AdaptiveBandit
- We Test it on multiple Simplified Potentials and in a Protein Folding Scenario
- We find that this Framework performs similarly to or better than previous Methods in every type of Test Potentials
- Furthermore, it Provides a Novel Framework to Develop new Sampling Algorithms with better Asymptotic Characteristics
My Own Short Summary
Put it Here
Paper Summary
- Why to use Computer Molecular Simulation
- Sampling in Molecular Simulations and Its Difficulties
- Markov State Models and multi-armed Bandit Relations
- Methods 2.1. Methods 2.2. Multi-armed Bandit Problem 2.3. AdaptiveBandit 2.4. Solving the Multi-armed Bandit Problem 2.5. AdaptiveBandit with Knowledge-based Initialization 2.6. Other Adaptive Sampling Algorithms 2.7. Langevin Dynamics on 2D Potentials 2.8. MD Simulation Setup
- Results 3.1. Performance Testing on 2D Potentials 3.2. Using System External Knowledge 3.3. Testing on Protein Folding Simulations
- Conclusions
Key Conclusion
put it here
Interesting Points
put it here
Disagreeing with Author
it’s possible!!
Re-Implementing Attempt(s) / Fusing into my Research
- Did You attempted to re-implement it?
- Are you Goning to Implement it in your own research?
- How?
- Can you plan it?
- Is it Possible?
Related Literature
Some of the interesting literature from either citated by or citations.