Ashutosh Chapagain
Papers
1
Total Citations
2
H-Index
1
About
Ashutosh Chapagain is a researcher advancing the frontier of planning under uncertainty, with a core focus on bridging the worlds of symbolic reasoning and deep reinforcement learning. His most notable contribution is the development of DiSProD (Differentiable Symbolic Propagation of Distributions for Planning), introduced in a 2023 paper. This innovative online planner tackles the notoriously difficult problem of handling probabilistic transitions within continuous state and action spaces. Rather than relying on purely numerical sampling, DiSProD constructs a symbolic graph that efficiently propagates the distribution of future trajectories conditioned on a policy, leveraging independence assumptions to maintain tractability. This approach offers a principled and differentiable alternative to traditional model-based planning methods, enabling more robust decision-making in complex, stochastic environments. While his work is still early in its citation lifecycle, the foundational nature of DiSProD marks Chapagain as a rising voice in the integration of symbolic AI with modern differentiable systems, promising significant impact on autonomous agents and robotics.
Research Focus
Key Achievements
Top Papers
- 1