Reinforcement Learning

Deep Reinforcement Learning for Adaptive Systems

Research Overview

Led by Dr. Deepali Salwan (PhD in Deep Reinforcement Learning), our RL research focuses on developing robust algorithms that can learn optimal behaviors in uncertain, dynamic environments with applications spanning from robotics to healthcare.

Key Focus Areas

Deep Q-Learning Architectures

Advanced value-based methods for complex state spaces

Policy Gradient Methods

Direct policy optimization for continuous control

Model-Based RL

Learning world models for sample-efficient planning

Multi-Task Transfer Learning

Knowledge sharing across related reinforcement learning tasks

Safe Reinforcement Learning

Constraint satisfaction and safety guarantees during exploration

Active Projects

SafeRL Framework: Industry-ready toolkit for deploying RL in safety-critical applications
BioAdaptive RL: Reinforcement learning for medical devices and rehabilitation
Sim-to-Real Transfer: Bridging simulation and physical deployment gaps