Eric Rosen
Papers
2
Total Citations
15
H-Index
2
About
Eric Rosen is an emerging researcher working at the intersection of robotics, artificial intelligence, and human activity recognition. His work focuses on two primary domains: leveraging large language models (LLMs) for intelligent robotic planning and developing deep learning architectures for complex human activity classification. Rosen's most notable contribution, **CAPE** (Corrective Actions from Precondition Errors), addresses a critical limitation in LLM-driven robotics — the inability to recover meaningfully from action failures. Rather than simply retrying failed actions, CAPE extracts commonsense knowledge from LLMs to diagnose and resolve the *underlying causes* of errors, representing a meaningful step forward in building truly adaptive robotic systems. This work has garnered 9 citations since its 2022 publication. His complementary work on **CHARM**, a hierarchical deep learning model for human activity recognition using motion sensors, tackles the challenge of classifying complex, multi-layered human behaviors — moving beyond traditional event-based recognition like step counting or fall detection. With 6 citations, this research demonstrates Rosen's broader interest in making machines better understand human behavior. Collectively, his contributions reflect a coherent research vision: building intelligent systems that can perceive, understand, and adapt to the complexities of human environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2