Papers
7
Total Citations
92
H-Index
5
About
Roma Patel is a researcher at the intersection of natural language processing and robotics, specializing in enabling robots to understand and execute complex, human-like instructions. Her work focuses on grounding natural language commands—particularly those with temporal and contextual constraints—into formal task specifications like Linear Temporal Logic (LTL). Patel’s major contributions include developing methods for robot object retrieval using contextual queries, allowing robots to locate objects based on nuanced descriptions rather than simple labels. She has also pioneered approaches to map natural language to “lifted LTL,” enabling generalization across new domains without task-specific supervision. Her most cited paper, “Robot Object Retrieval with Contextual Natural Language Queries” (43 citations), demonstrates how robots can interpret commands like “bring me the red cup from the kitchen” by leveraging environmental context. Other notable works address affordance-based retrieval and hierarchical planning with state abstractions for non-Markovian tasks. Patel’s research has been cited over 90 times, reflecting its impact on making human-robot interaction more intuitive and flexible. Her recent exploration of child-robot interaction in performative art settings highlights her commitment to broadening robotics’ social applications.
Research Focus
Key Achievements
Top Papers
- 1Robot Object Retrieval with Contextual Natural Language Queries43 citations · 2020
- 2
- 3Generalizing to New Domains by Mapping Natural Language to Lifted LTL13 citations · 2022
- 4Affordance-based robot object retrieval7 citations · 2021
- 5
- 6
- 7Planning with State Abstractions for Non-Markovian Task Specifications2 citations · 2019