Papers
42
Total Citations
3,025
H-Index
20
About
Pete Florence is a pioneering robotics and AI researcher whose work sits at the intersection of robot learning, computer vision, and large-scale foundation models. He is perhaps best known for bridging the gap between modern language and vision models and real-world robotic control, with landmark contributions including "Code as Policies" (561 citations), which demonstrated that code-generating language models can be repurposed to write executable robot policy programs from natural language commands, and "RT-2" (267 citations), which showed that vision-language models trained on internet-scale data can transfer emergent semantic reasoning directly into robotic action. His involvement in "PaLM-E" (350 citations) further established embodied multimodal language models as a viable framework for grounded real-world inference. Florence also made foundational contributions to 3D shape representation through "DeepSDF" (259 citations) and to robot perception through "Dense Object Nets" (100 citations), introducing dense visual descriptors that enabled task-agnostic object understanding for manipulation. His broader portfolio spans deformable object manipulation, keypoint-based affordances, and compositional multimodal reasoning. Collectively, his research has amassed well over 2,300 citations, reflecting profound and growing influence on how intelligent robots perceive, reason about, and act in the physical world.
Research Focus
Key Achievements
Top Papers
- 1Code as Policies: Language Model Programs for Embodied Control561 citations · 2023
- 2PaLM-E: An Embodied Multimodal Language Model350 citations · 2023
- 3RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control267 citations · 2023
- 4DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation259 citations · 2019
- 5Inner Monologue: Embodied Reasoning through Planning with Language Models206 citations · 2022
- 6Socratic Models: Composing Zero-Shot Multimodal Reasoning with Language171 citations · 2022
- 7KPAM: KeyPoint Affordances for Category-Level Robotic Manipulation170 citations · 2022
- 8
- 9
- 10