Biswadip Dey
Papers
5
Total Citations
318
H-Index
4
About
Biswadip Dey is a researcher whose work bridges the frontiers of computational ecology and physics-informed machine learning. His most impactful research explores how environmental stressors, particularly neonicotinoid pesticides, disrupt the intricate social fabric of bumblebee colonies. In a landmark 2018 study cited over 290 times, Dey used an innovative automated robotic platform to reveal that pesticide exposure degrades nest behavior, social networks, and thermoregulation—providing a mechanistic link between sublethal exposure and colony decline. This work has become foundational for understanding pollinator health and conservation. Simultaneously, Dey is advancing the field of scientific machine learning. He has developed differentiable contact models that extend Lagrangian and Hamiltonian neural networks, enabling these architectures to accurately learn hybrid dynamics involving impacts and discontinuities. This work, published in 2021, introduces crucial inductive biases for modeling physical systems where energy conservation and contact events coexist. By combining rigorous behavioral ecology with cutting-edge differentiable physics, Dey’s research demonstrates a rare ability to draw insights from nature while building the computational tools to simulate it. His work on collective motion and cyclic pursuit further underscores his commitment to understanding and synthesizing complex, multi-agent behaviors.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Collective motion under beacon-referenced cyclic pursuit10 citations · 2018
- 4
- 5Reconstruction, Analysis and Synthesis of Collective Motion2 citations · 2015