Ravi Mangal
Papers
2
Total Citations
13
H-Index
2
About
Ravi Mangal is a researcher working at the intersection of formal methods, autonomous systems, and deep learning, with a focus on making AI-driven systems safe and reliable. His most recognized contribution is the development of **DeepDECS**, a novel framework for synthesizing correct-by-construction controllers for autonomous systems that rely on deep neural network (DNN) classifiers for perception-based decision-making. This work addresses one of the most pressing challenges in modern AI deployment: despite the impressive capabilities of deep learning, providing formal safety guarantees for systems that incorporate DNNs remains extremely difficult. DeepDECS bridges this gap by integrating discrete-event control synthesis with probabilistic models of DNN classifier behavior, enabling the design of controllers that provably satisfy safety specifications even in the presence of perception uncertainty. Mangal's work has appeared across multiple publication venues, with his 2024 journal extension of DeepDECS accumulating 10 citations and the 2022 conference version garnering 3 citations. His research is particularly valuable for students and engineers working on safety-critical autonomous systems, offering rigorous, tool-supported methods for trustworthy AI integration.
Research Focus
Key Achievements
Top Papers
- 1
- 2