Papers
14
Total Citations
205
H-Index
7
About
David Paulius is a leading researcher in knowledge representation and task planning for service robotics, with a particular focus on enabling robots to perform complex manipulation tasks like cooking. His most influential work, the survey "A Survey of Knowledge Representation in Service Robotics" (99 citations), established a foundational framework for the field. Paulius is best known for developing the **Functional Object-Oriented Network (FOON)** , a knowledge graph representation that captures the functional relationships between objects and actions. This innovation, detailed in papers like "Long-Horizon Planning and Execution With Functional Object-Oriented Networks" (11 citations), allows robots to reason about tasks symbolically and adapt plans to novel situations. He also introduced a manipulation taxonomy that disambiguates motion types and consolidates aliases, paving the way for skill transfer between tasks. More recently, Paulius has explored using Large Language Models (LLMs) for error recovery in robot planning, as seen in his work on **CAPE** (Corrective Actions from Precondition Errors, 13 citations), which enables robots to diagnose and resolve action failures rather than simply retrying them. His research consistently bridges symbolic reasoning with real-world robotic execution, making him a key figure in advancing autonomous manipulation.
Research Focus
Key Achievements
Top Papers
- 1A Survey of Knowledge Representation in Service Robotics99 citations · 2019
- 2Manipulation Motion Taxonomy and Coding for Robots21 citations · 2019
- 3Approximate Task Tree Retrieval in a Knowledge Network for Robotic Cooking17 citations · 2022
- 4Task Planning with a Weighted Functional Object-Oriented Network15 citations · 2021
- 5
- 6Long-Horizon Planning and Execution With Functional Object-Oriented Networks11 citations · 2023
- 7
- 8
- 9
- 10