Learning Clustered Representation for Complex Free Energy Landscapes
This work addresses the challenge of representing complex free energy landscapes in fields like computational chemistry or biophysics, offering a scalable and unsupervised approach without requiring predefined metrics or cluster numbers, though it appears incremental as it builds on existing dimensionality reduction and clustering methods.
The paper tackled the problem of analyzing complex free energy landscapes (FEL) by developing an unsupervised method, Information Distilling of Metastability (IDM), to produce reduced and clustered representations, achieving physically meaningful partitions into metastable states for downstream tasks.
In this paper we first analyzed the inductive bias underlying the data scattered across complex free energy landscapes (FEL), and exploited it to train deep neural networks which yield reduced and clustered representation for the FEL. Our parametric method, called Information Distilling of Metastability (IDM), is end-to-end differentiable thus scalable to ultra-large dataset. IDM is also a clustering algorithm and is able to cluster the samples in the meantime of reducing the dimensions. Besides, as an unsupervised learning method, IDM differs from many existing dimensionality reduction and clustering methods in that it neither requires a cherry-picked distance metric nor the ground-true number of clusters, and that it can be used to unroll and zoom-in the hierarchical FEL with respect to different timescales. Through multiple experiments, we show that IDM can achieve physically meaningful representations which partition the FEL into well-defined metastable states hence are amenable for downstream tasks such as mechanism analysis and kinetic modeling.