Research Overview
The three current areas of research in CLAN Labs are:
Reinforcement Learning
The current themes are
- Regret/Sample Complexity Guarantees for Discounted/Average Reward MDPs
- Constrained Markov Decision Process
- Linearizable Optimization with Application to DR Submodular Bandits
- Combinatorial Submodular Bandits
- Multi-Agent Reinforcement Learning
Foundation Models and Generative AI
The current themes are
- Reinforcement Learning with Human Feedback
- Prompt Optimization for LLM Alignment
- Guarantees for Generative AI
Applications of Machine Learning
The current themes are
- Generative Models for Drug Discovery
- Foundation Models for Genomics
- Application of Machine Learning tools for healthcare (e.g., Sepsis)
- Network Resource Allocation, including Vehicular Networks
- Applications to Transportation Networks (e.g., Traffic Signal Control, Lane Changing, Ridesharing)
Quantum Machine Learning
The current themes are:
- Quantum Reinforcement Learning
- Variational Quantum Circuits for Machine Learning
- Quantum Computing for Imaging Applications
- Grover's Search