Abhinav Shirivastava
Papers
1
Total Citations
3
H-Index
1
About
Abhinav Shirivastava is pioneering advances at the intersection of robot learning and visuo-spatial generalization, with a focus on building policies that transfer reliably across diverse environments and object instances. His most cited work, "P3-PO: Prescriptive Point Priors for Visuo-Spatial Generalization of Robot Policies" (2025), introduces a novel framework that leverages prescriptive point priors to guide policy behavior, enabling robots to robustly adapt to novel spatial configurations and visual conditions without requiring exhaustive retraining. This contribution addresses a critical bottleneck in robot learning—the fragility of policies trained on limited data—by embedding geometric and semantic priors directly into the policy architecture. Shirivastava’s research sits at the nexus of computer vision, robotics, and machine learning, offering practical pathways toward more generalizable and sample-efficient robot systems. His work has already garnered attention in the community, with early citations reflecting its relevance to ongoing efforts in scalable robot learning. By tackling the challenge of out-of-distribution generalization, Shirivastava is helping to move robot learning from controlled labs toward real-world deployment, where variability is the norm.
Research Focus
Key Achievements
Top Papers
- 1