Bill DeRose

Berkeley College, University of California, Berkeley

Papers

Total Citations

603

H-Index

About

Bill DeRose is a leading researcher in robotic manipulation, with a primary focus on enabling robots to grasp and handle objects with human-like dexterity and reliability. His work centers on the grand challenge of universal picking—the ability for robots to reliably grasp a diverse range of novel objects from cluttered heaps, a critical capability for e-commerce, manufacturing, and home service robotics. DeRose’s most impactful contribution is his pioneering research on learning ambidextrous robot grasping policies, published in 2019, which has garnered 578 citations. This work addresses the inherent uncertainty in robotic grasping by developing algorithms that optimize rate, reliability, and range, marking a significant leap forward in the field. He also developed Dex-Net as a Service (DNaaS), a cloud-based grasp planning system that democratizes access to advanced manipulation software, reducing infrastructure overhead for automation systems. DeRose’s research has profound implications for industries seeking to automate complex tasks, and his highly cited work underscores his influence in advancing the state of the art in robotic manipulation.