Publications

2026

• Decoupled Continuous-Time Reinforcement Learning via Hamiltonian Flow
  Minh Nguyen
  [Preprint 2026] · [Paper] · [Website] · [Code]
  

  This work studies reinforcement learning in continuous time, where standard discrete-time action-value updates become less informative as time steps shrink. It proposes a decoupled framework built around value flow and Hamiltonian structure.

• GRL-SNAM: Geometric Reinforcement Learning with Differential Hamiltonians for Navigation and Mapping in Unknown Environments
  Aditya Sai Ellendula, Yi Wang, Minh Nguyen, Chandrajit Bajaj
  [ICLR 2026] · [Paper] · [Website] · [Code]
  

  This work studies navigation and mapping in unknown environments through a differential reinforcement learning formulation built around adaptive Hamiltonian structures.

2025

• A Differential and Pointwise Control Approach to Reinforcement Learning
  Minh Nguyen and Chandrajit Bajaj
  [NeurIPS 2025] · [Paper] · [Website] · [Code]
  

  This work develops a reinforcement learning theory centered on differential and pointwise control structure, aiming for stronger physics-informed behavior and more reliable learning.

• Learning Generalized Hamiltonian Dynamics with Stability from Noisy Trajectory Data
  Luke McLennan, Yi Wang, Ryan Farell, Minh Nguyen, Chandrajit Bajaj
  [Preprint 2025] · [Paper] · [Code]
  

  This work introduces a robust framework for learning various generalized Hamiltonian dynamics from noisy, sparse phase-space data and in an unsupervised manner based on variational Bayesian inference.

• Stochastic Differential Policy Optimization: A Rough Path Approach to Reinforcement Learning
  Minh Nguyen and Chandrajit Bajaj
  [TASC, COLT 2025] · [Paper]
  

  This work develops a stochastic-differential view of policy optimization, combining ideas from stochastic control and rough path theory to study reinforcement learning in continuous-time.

• Knowledge-Enhanced Framework for Biomedical Entity and Relation Extraction
  Minh Nguyen and Phuong Le
  [LNCS, Springer, 2025] · [Paper] · [Code]
  

  This work proposes a knowledge-enhanced framework for biomedical entity and relation extraction, combining domain structure with machine learning to improve information extraction quality.

• Reinforcement Learning for Molecular Dynamics Optimization: A Stochastic Pontryagin Maximum Principle Approach
  Chandrajit Bajaj, Minh Nguyen, and Conrad Li
  [CCIS, Springer, 2025] · [Paper]
  

  This work presents a novel reinforcement learning framework designed to optimize molecular dynamics. Through extensive experimentation on six distinct molecules, including Bradykinin and Oxytocin, we demonstrate competitive performance against other unsupervised physics-based methods.

• Low-cost Robust Night-time Aerial Material Segmentation through Hyperspectral Data and Sparse Spatio-Temporal Learning
  Chandrajit Bajaj, Minh Nguyen, and Shubham Bhardwaj
  [CCIS, Springer, 2025] · [Paper] · [Code]
  

  This work proposes an innovative Siamese framework that uses time series-based compression to effectively perform material segmentation on high spectral data under poor lighting and atmospheric conditions.

2024

• Motion Code: Robust Time series Classification and Forecasting via Sparse Variational Multi-Stochastic Processes Learning
  Minh Nguyen and Chandrajit Bajaj
  [Preprint 2024] · [Paper] · [Code]
  

  This work proposes a novel framework that views each time series, regardless of length, as a realization of a continuous-time stochastic process. It also introduces the concept of "most informative timestamps" to infer a sparse approximation of the individual dynamics from these vectors and fully captures diverse underlying dynamics in an integrated manner, enabling simultaneous classification and forecasting of time series.

• Physics-Informed Neural Networks via Stochastic Hamiltonian Dynamics Learning
  Chandrajit Bajaj and Minh Nguyen
  [LNNS, Springer, 2024] · [Paper] · [Code]
  

  This work proposes novel learning frameworks called NeuralPMP to tackle optimal control problems through the Pontryagin maximum principle and the reduced Hamiltonian networks.

2020

• Convex function through Doob-Meyer decomposition
  Minh Nguyen
  [Preprint 2020] · [Paper]
  This work studies convex functions through Brownian motion and the Doob–Meyer decomposition. It derives a strong form of Ito's lemma for convex functions and uses probabilistic tools to recover second-order differentiability.

2017

• Irredundant generating sets and dimension-like invariants of the finite group
  Minh Nguyen
  [Preprint 2017] · [Paper]
  

  This work classifies all irredundant generating sets of the symmetry permutation group and derive further interesting properties for the alternating group and calculate dimension-like invariants of wreath products.

2016

• Ideals in the Goldman Algebra
  Minh Nguyen
  [Preprint 2016] · [Paper]
  This work studies the ideals of the Goldman Lie algebra. We construct an algebra homomorphism from the original structure to a simpler commutative algebraic structure. From here, we discover a non-trivial infinite class of ideals in this Lie algebra.