Yash Shukla
Papers
3
Total Citations
19
H-Index
2
About
Yash Shukla is a robotics and artificial intelligence researcher whose work spans haptic perception, continual learning, and the integration of large language models with reinforcement learning. His most cited contribution, "Haptic Knowledge Transfer Between Heterogeneous Robots using Kernel Manifold Alignment" (2020, 14 citations), demonstrates his early focus on enabling robots to understand object properties through non-visual sensory modalities — a meaningful step toward more human-like robotic perception. By developing methods for transferring haptic knowledge across different robot platforms, Shukla addressed a critical challenge in making tactile learning scalable and generalizable. His more recent research reflects a compelling evolution toward open-world autonomy. In "A Framework for Neurosymbolic Goal-Conditioned Continual Learning" (2024), he bridges symbolic planning and reinforcement learning to help agents adapt to unpredictable environments — a frontier challenge in autonomous systems. His work on LLM-generated sub-goals for reinforcement learning agents further positions him at the exciting intersection of foundation models and robot decision-making. Collectively, Shukla's research addresses some of the most pressing questions in embodied AI: how machines can learn, adapt, and reason in complex, ever-changing worlds, making him a researcher worth following as these fields rapidly mature.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3