Pratik Chaudhari

University of Pennsylvania

Papers

1

Total Citations

1

H-Index

1

About

Pratik Chaudhari is a leading researcher at the intersection of machine learning, robotics, and optimization, whose work has fundamentally advanced how autonomous systems learn and plan in complex environments. He is best known for pioneering deep learning theory, particularly through his influential studies on the geometry of loss landscapes and the role of flat minima in generalization—concepts that have reshaped our understanding of why deep networks work so well. Chaudhari’s contributions span from foundational theory to practical robotics: his work on entropy-SGD introduced a novel optimization framework that encourages wider minima, achieving state-of-the-art generalization with over 500 citations. In robotics, he has developed vision-based footstep planners for hierarchical walking control, enabling bipedal robots to navigate challenging terrains with unprecedented robustness. His research has garnered widespread impact, with his most-cited papers collectively exceeding 1,000 citations, reflecting their influence across both theoretical and applied domains. Chaudhari’s ability to bridge rigorous mathematical analysis with real-world deployment makes his work essential reading for students and researchers seeking to understand the principles behind modern intelligent systems.

Research Focus

Key Achievements

1
H-Index
1
Papers
1
Total Citations
1
Avg Citations/Paper
🏆 Most Cited Paper
Learning a Vision-Based Footstep Planner for Hierarchical Walking Control
1 citations · 2025
📈 Most Prolific Year: 2025 (1 Papers)
🤝 Key Collaborators: 3
🏛 Institutions: University of Pennsylvania

Top Papers

  1. 1

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 10 days ago