Jyotirmoy V. Deshmukh
Papers
10
Total Citations
74
H-Index
4
About
Jyotirmoy V. Deshmukh is a prominent researcher at the intersection of formal methods, robotics, and autonomous systems, with a particular focus on using temporal logic to make learning-based control safer, more robust, and interpretable. His most influential work centers on integrating Signal Temporal Logic (STL) into reinforcement learning frameworks — enabling robots to learn from demonstrations while adhering to rigorous safety specifications, rather than relying on hand-crafted or easily exploitable reward functions. His 2021 paper on learning from demonstrations using STL has garnered 26 citations, reflecting strong community interest in bridging formal verification with machine learning. Deshmukh has also made notable contributions to adversarial testing and falsification of cyber-physical systems, including autonomous vehicles and aircraft, developing reinforcement learning-based approaches to automatically discover system failures. His more recent work expands into natural language interfaces for STL specification and spatio-temporal logic for multi-agent distributed systems, demonstrating a broad and forward-looking research vision. As a guest editor for a special issue on autonomous systems design, he has further shaped scholarly discourse in this rapidly evolving field, establishing himself as a key voice in safe and verifiable autonomy.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Model-based Reinforcement Learning from Signal Temporal Logic Specifications14 citations · 2020
- 4
- 5
- 6Poster Abstract: Learning from Demonstrations with Temporal Logics3 citations · 2022
- 7Learning from Demonstrations using Signal Temporal Logic3 citations · 2021
- 8
- 9Guest Editors’ Introduction: Special Issue on Autonomous Systems Design1 citations · 2022
- 10