2024, JCTC, AdaptiveBandit: A Multi-armed Bandit Framework for Adaptive Sampling in Molecular Simulations

[!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


  1. Introduction
  • Why to use Computer Molecular Simulation
  • Sampling in Molecular Simulations and Its Difficulties
  • Markov State Models and multi-armed Bandit Relations
  1. 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
  2. Results 3.1. Performance Testing on 2D Potentials 3.2. Using System External Knowledge 3.3. Testing on Protein Folding Simulations
  3. 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.