Subham Banga

Papers

Total Citations

H-Index

About

Subham Banga is a researcher advancing the frontiers of human-robot interaction and transfer learning. His work focuses on enabling robots to intuitively anticipate and support human actions, particularly in complex, real-world environments. Banga’s most-cited paper, “Selecting Source Tasks for Transfer Learning of Human Preferences” (2024, 2 citations), tackles a critical challenge: how to efficiently transfer human behavioral preferences from simpler, well-understood tasks to more intricate target scenarios. By developing methods to select optimal source tasks, he reduces the need for extensive human demonstrations in every new context, making proactive robotic assistance more practical and scalable. This contribution is foundational for creating robots that can learn and adapt to individual user preferences without exhaustive retraining. Banga’s research sits at the intersection of machine learning, cognitive science, and robotics, with implications for assistive technologies, autonomous systems, and personalized AI. His work is particularly notable for addressing the “cold start” problem in preference learning, offering a principled approach to leveraging prior knowledge. As the field moves toward more autonomous and socially aware robots, Banga’s insights into efficient preference transfer will be instrumental in shaping how machines learn to collaborate with humans seamlessly.