Papers

5

Total Citations

723

H-Index

5

About

Marco Cusumano-Towner is a leading researcher at the intersection of robotics, computer vision, and probabilistic programming. His early work revolutionized robotic manipulation of deformable objects, most notably through his highly cited paper on cloth grasp point detection (429 citations), which introduced a geometric-cue-based algorithm enabling reliable robotic towel folding. He further advanced this domain by developing hidden Markov model approaches for bringing clothing into desired configurations using limited perception (153 citations). Cusumano-Towner's most transformative contribution is the creation of Gen, a general-purpose probabilistic programming system with programmable inference (128 citations). This system addresses critical limitations in existing probabilistic programming frameworks, offering the flexibility and efficiency needed for challenging real-world applications in computer vision and robotics. His research also explores probabilistic programs for inferring the goals of autonomous agents, bridging robotics and cognitive science, and he developed AIDE, an algorithm for measuring the accuracy of probabilistic inference systems. Through these contributions, Cusumano-Towner has fundamentally advanced both the theoretical foundations and practical capabilities of intelligent robotic systems and probabilistic computing.

Research Focus

Key Achievements

5
H-Index
5
Papers
723
Total Citations
145
Avg Citations/Paper
🏆 Most Cited Paper
Cloth grasp point detection based on multiple-view geometric cues with application to robotic towel folding
429 citations · 2010
📈 Most Prolific Year: 2017 (2 Papers)
🤝 Key Collaborators: 11
🏛 Institutions: University of California, Berkeley, Massachusetts Institute of Technology

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 12 days ago