Research Projects

My research develops learning and optimization methods that connect foundational algorithmic advances with scientific and healthcare applications. Current projects focus on the following three areas.

1. Communication-Efficient Federated Learning

This project studies federated learning methods that substantially reduce communication overhead while preserving convergence and learning performance. We develop dimension-free algorithms that exchange only a constant number of scalar values, together with Hessian-informed acceleration and participation-aware methods for heterogeneous and intermittently available devices. Applications include edge intelligence, wireless networks, and distributed fine-tuning of large language models.

2. Multi-Objective Learning for Science

This project develops learning and optimization methods for problems with multiple, potentially competing objectives. We study stochastic multi-gradient algorithms, multi-objective reinforcement learning, and federated multi-objective learning, with an emphasis on convergence, sample efficiency, and robust trade-off discovery. These methods are designed for scientific workflows in which simulations and experiments are costly and observations are limited.

3. AI for Healthcare

This project develops AI methods for healthcare and biomedicine, with an emphasis on protein therapeutics. Our current work uses multi-objective learning to design nanobody sequences that balance affinity, stability, selectivity, and solubility. By coupling predictive models with wet-lab feedback, this research aims to reduce experimental trial and error and accelerate therapeutic development for cancer, autoimmune disorders, and infectious diseases.

Awards

  1. NIH/NIGMS R16 Award: Optimizing Nanobody Sequence Design through Multi-Objective Engineering. Principal Investigator, Award No. R16GM159671, 2025–2029.
  2. NSF Award: Distributed Optimization with Dimension-Free Communication: Theory and Algorithms. Principal Investigator, Award No. 2537471, 2026–2029.
  3. NVIDIA Academic Grant: Federated Edge AI for Personalized In-Home Heart Failure Monitoring. Award Recipient, 2026.