Yeli Feng

Papers

Total Citations

H-Index

About

Yeli Feng is a researcher at the forefront of making machine learning safe and reliable for real-world, safety-critical systems. Her work focuses on the intersection of embedded systems, cyber-physical systems (CPS), and robust machine learning, with a particular emphasis on out-of-distribution (OOD) detection. Feng’s major contribution lies in bridging the gap between theoretical ML robustness and practical deployment constraints. Her most cited paper, “Embedded out-of-distribution detection on an autonomous robot platform” (2021, 13 citations), demonstrates a pioneering approach to implementing OOD detection directly on resource-constrained autonomous robots, addressing a critical vulnerability in modern CPS. This work is notable for moving beyond simulation to real hardware, showing how to maintain reliability when testing data deviates from training distributions—a key challenge in safety-critical applications like autonomous driving and industrial robotics. By tackling the “last mile” of ML deployment, Feng’s research has significant implications for trustworthy AI in embedded environments, earning her recognition as a rising voice in dependable autonomous systems.