About me [causality, bayes, reinforcement learning, software engineering, optimization and control]

With transdisciplinary expertise in statistics, software engineering, (multi-objective) optimization and control theory, my research perspective lies in experimental design, data driven control and sequential decision making for precision medicine and robotics applications.

To achieve this, I have dipped into the following research activities:

  • Data driven decision and control
    • Multi-objective feedback optimization and adaptive control with uncertain plant models (e.g. deep neural networks) M-HOF-Opt
    • Multi-level optimization with hypothesis posterior for multi-modal space optimization using RL and Bayesian Optimization ReinBo
    • Optimal path learning for multi-parameter persistence homology (Toplogical Data Analysis) persistence homology learning
    • Understanding inference for control with Probabilistic Graphical Model Tutorial Paper
    • Exploration in reinforcement learning with sparse reward maxEntropy, ReinBo
    • High dimensional Robust Model Predicative Control tubeMPC
  • Out of distribution generalization
    • Open source modular software design for domain generalization and multi-modal deep learning DomainLab
    • Causal and variational inference based domain generalization for deep learning HDUVA
    • Multi-objective Bayesian optimization for distributed model selection under domain shift RFMS
    • Bayesian neural network does not generalize paper
    • Approximate causal discover with hidden confounders
  • Learning sequential data and dynamic systems
    • Learning based non-linear observer design for dynamic systems kkl observer
    • Learning Opinion dynamics IFAC paper
    • Port-Hamiltonian modeling, learning and controlĀ (e.g. bio-chemical reaction networks)
    • Generalized Additive Model for Functional Data Analysis mlrFDA

For more research activities, see my google scholar.

I am also an open source software developer (R, python, cpp, java) for machine learning and reinforcement learning see my github.

Since 2021, I have been a continuous reviewer for AI conferences like ICML, ICLR, NeuroIPS.

For system, optimization and control theory, I am trained in

  • formal method, nonlinear control
  • nonlinear contraint optimization, LMI
  • networked optimization and control, opinion dynamics
  • model predictive control
  • power system optimization and control
  • infinite dimensional control (PDE)
  • (soft) robotics

I value the role of algebraic statistics and differential geometry in my research

During my free time, I like exploring the nature with friends, enjoy various sports (e.g. table tennis, badminton, swimming, entry level martial arts).