Resume

• 2014

  Physics

  University of Illinois at Urbana-Champaign

• 2010

  Bachelor of Science (B.S.)

  Physics

  Nanjing University

• Research

  Understanding of physical systems depends crucially on sufficient sampling of system conformations. However

  the presence of energy barrier generally impedes efficient sampling. One way to solve this issue is to add biasing forces to “pull” the system...

  Research

  Data Science

  Data Analysis

  Molecular Dynamics

  C

  Version Control Tools

  Computational Chemistry

  Neural Networks

  Bash

  LaTeX

  Deep Learning

  Version Control

  Linux

  MATLAB

  C++

  CUDA

  Computational Physics

  Machine Learning

  University Teaching

  Python

  Collective variable discovery and enhanced sampling using autoencoders: Innovations in network architecture and error function design

  Auto-associative neural networks (“autoencoders”) present a powerful nonlinear dimensionality reduction technique to mine data-driven collective variables from molecular simulation trajectories. This technique furnishes explicit and differentiable expressions for the nonlinear collective variables

  making it ideally suited for integration with enhanced sampling techniques for accelerated exploration of configurational space. In this work

  we describe a number of sophistications of the neural network architectures to improve and generalize the process of interleaved collective variable discovery and enhanced sampling. We employ circular network nodes to accommodate periodicities in the collective variables

  hierarchical network architectures to rank-order the collective variables

  and generalized encoder-decoder architectures to support bespoke error functions for network training to incorporate prior knowledge. We demonstrate our approach in blind collective variable discovery and enhanced sampling of the configurational free energy landscapes of alanine dipeptide and Trp-cage using an open-source plugin developed for the OpenMM molecular simulation package.

  Collective variable discovery and enhanced sampling using autoencoders: Innovations in network architecture and error function design

  Nonlinear discovery of slow molecular modes using state-free reversible VAMPnets

  The success of enhanced sampling molecular simulations that accelerate along collective variables (CVs) is predicated on the availability of variables coincident with the slow collective motions governing the long-time conformational dynamics of a system. It is challenging to intuit these slow CVs for all but the simplest molecular systems

  and their data-driven discovery directly from molecular simulation trajectories has been a central focus of the molecular simulation community to both unveil the important physical mechanisms and drive enhanced sampling. In this work

  we introduce state-free reversible VAMPnets (SRV) as a deep learning architecture that learns nonlinear CV approximants to the leading slow eigenfunctions of the spectral decomposition of the transfer operator that evolves equilibrium-scaled probability distributions through time. Orthogonality of the learned CVs is naturally imposed within network training without added regularization. The CVs are inherently explicit and differentiable functions of the input coordinates making them well-suited to use in enhanced sampling calculations. We demonstrate the utility of SRVs in capturing parsimonious nonlinear representations of complex system dynamics in applications to 1D and 2D toy systems where the true eigenfunctions are exactly calculable and to molecular dynamics simulations of alanine dipeptide and the WW domain protein.

  Nonlinear discovery of slow molecular modes using state-free reversible VAMPnets

  Promoting transparency and reproducibility in enhanced molecular simulations

  The PLUMED consortium unifies developers and contributors to PLUMED

  an open-source library for enhanced-sampling

  free-energy calculations and the analysis of molecular dynamics simulations. Here

  we outline our efforts to promote transparency and reproducibility by disseminating protocols for enhanced-sampling molecular simulations.

  Promoting transparency and reproducibility in enhanced molecular simulations

  High-resolution Markov state models for the dynamics of Trp-cage miniprotein constructed over slow folding modes identified by state-free reversible VAMPnets

  High-resolution Markov state models for the dynamics of Trp-cage miniprotein constructed over slow folding modes identified by state-free reversible VAMPnets

  Molecular enhanced sampling with autoencoders: On‐the‐fly collective variable discovery and accelerated free energy landscape exploration

  Macromolecular and biomolecular folding landscapes typically contain high free energy barriers that impede efficient sampling of configurational space by standard molecular dynamics simulation. Biased sampling can artificially drive the simulation along prespecified collective variables (CVs)

  but success depends critically on the availability of good CVs associated with the important collective dynamical motions. Nonlinear machine learning techniques can identify such CVs but typically do not furnish an explicit relationship with the atomic coordinates necessary to perform biased sampling. In this work

  we employ auto‐associative artificial neural networks (“autoencoders”) to learn nonlinear CVs that are explicit and differentiable functions of the atomic coordinates. Our approach offers substantial speedups in exploration of configurational space

  and is distinguished from existing approaches by its capacity to simultaneously discover and directly accelerate along data‐driven CVs. We demonstrate the approach in simulations of alanine dipeptide and Trp‐cage

  and have developed an open‐source and freely available implementation within OpenMM.

  Molecular enhanced sampling with autoencoders: On‐the‐fly collective variable discovery and accelerated free energy landscape exploration

  Capabilities and Limitations of Time-lagged Autoencoders for Slow Mode Discovery in Dynamical Systems

  Time-lagged autoencoders (TAEs) have been proposed as a deep learning regression-based approach to the discovery of slow modes in dynamical systems. However

  a rigorous analysis of nonlinear TAEs remains lacking. In this work

  we discuss the capabilities and limitations of TAEs through both theoretical and numerical analyses. Theoretically

  we derive bounds for nonlinear TAE performance in slow mode discovery and show that in general TAEs learn a mixture of slow and maximum variance modes. Numerically

  we illustrate cases where TAEs can and cannot correctly identify the leading slowest mode in two example systems: a 2D \"Washington beltway\" potential and the alanine dipeptide molecule in explicit water. We also compare the TAE results with those obtained using state-free reversible VAMPnets (SRVs) as a variational-based neural network approach for slow modes discovery

  and show that SRVs can correctly discover slow modes where TAEs fail.\n

  Capabilities and Limitations of Time-lagged Autoencoders for Slow Mode Discovery in Dynamical Systems

  Molecular Enhanced Sampling with Autoencoders

  Slow Molecular Modes Discovery with Machine Learning Methods

  Chen

  Wei

  Chen

  Qulab Inc.

  University of Illinois at Urbana-Champaign

  University of Chicago

  Greater Chicago Area

  * Derived and proved theoretical bounds for time-lagged autoencoders for stationary processes

  providing theoretical insights for future machine learning method development\n\n* Created a deep learning based model to extract hierarchical slow modes from time-series data (https://github.com/hsidky/srv)

  achieving O(N) space and O(N) time complexity

  as opposed to O(N^2) space and O(N^3) time complexity required by the state-of-the-art kernel based model

  Visiting PHD Researcher

  University of Chicago

  Greater Los Angeles Area

  * Built machine learning-based adaptive sampling and Markov model analysis pipelines

  successfully identifying sampling bottlenecks and help computational drug design

  Research Intern

  Qulab Inc.

  Urbana-Champaign

  Illinois Area

  * Taught lab and discussion sections for university physics courses for more than 400 engineering students in total

  awarded ``Excellent Teacher'' 4 times out of 6 semesters\n\n* Served as the mentor teaching assistant for one semester

  organized peer evaluation and offered post-lecture feedback for new TAs

  Graduate Teaching Assistant

  University of Illinois at Urbana-Champaign

  Urbana-Champaign

  Illinois Area

  * Applied autoencoders with data augmentation to discover important collective variables for molecular systems\n\n* Developed an enhanced sampling framework for molecular simulations (https://github.com/weiHelloWorld/accelerated_sampling_with_autoencoder)

  enabling orders of magnitude faster sampling than conventional techniques\n\n* Built a neural network based force plugin for a simulation package with native CUDA support (https://github.com/weiHelloWorld/ANN_Force) and\ncontributed its variant to a general-purpose package widely used by simulation community (https://github.com/weiHelloWorld/plumed2)

  PHD Researcher

  University of Illinois at Urbana-Champaign

  English

  Chinese

  Kuck Computational Science & Engineering Scholarship

  Department of Computational Science and Engineering at UIUC