Adarsh Sehgal

About

Adarsh Sehgal is a researcher specializing in deep reinforcement learning (RL) and evolutionary computation, with a particular focus on automated hyperparameter optimization and robotic manipulation. His work sits at the intersection of machine learning and robotics, addressing one of the most persistent challenges in deploying RL systems: the sensitivity of learning algorithms to parameter selection. Sehgal's most influential contribution, "Deep Reinforcement Learning Using Genetic Algorithm for Parameter Optimization" (2019), has garnered 106 citations and established him as a pioneer in applying genetic algorithms (GA) to automate the optimization of RL hyperparameters — a process traditionally dependent on costly manual tuning. This foundational work paved the way for his subsequent research, where he extended these ideas to robotic manipulation tasks using advanced architectures such as Deep Deterministic Policy Gradient (DDPG) combined with Hindsight Experience Replay (HER), demonstrated in his 2022 publications. Collectively, his research offers practical frameworks that make reinforcement learning more efficient and accessible for real-world robotics applications. With over 130 citations across his key works, Sehgal's contributions are shaping how autonomous systems are trained, making him a noteworthy voice in intelligent robotics and adaptive learning systems.

Research Focus

Key Achievements

Top Papers

1. 1
   Deep Reinforcement Learning Using Genetic Algorithm for Parameter Optimization

   106 citations · 2019
2. 2
   GA+DDPG+HER: Genetic Algorithm-Based Function Optimizer in Deep Reinforcement Learning for Robotic Manipulation Tasks

   16 citations · 2022
3. 3
   Automatic Parameter Optimization Using Genetic Algorithm in Deep Reinforcement Learning for Robotic Manipulation Tasks

   9 citations · 2022

Key Collaborators

Contact & Links