Max Robotics — Global Service Robot Directory + US Certification

About

Guanpeng Li is a leading researcher at the intersection of machine learning reliability and robotic vehicle security. His work focuses on ensuring that ML-driven systems—especially those in safety-critical domains like autonomous vehicles and industrial robotics—can withstand both accidental faults and malicious attacks. Li’s most influential contribution is **TensorFI**, a configurable fault injector for TensorFlow applications (63 citations), which provides a foundational tool for testing the resilience of ML models to hardware and software errors. He also pioneered **PID-Piper** (41 citations), a novel framework that enables robotic vehicles to recover from physical sensor attacks, such as tampering and spoofing, by actively correcting erroneous measurements. Building on this, his recent **DeLorean** framework (2024) unifies attack detection, diagnosis, and recovery, advancing the state of the art in securing autonomous systems. Through his position paper on error-resilient ML (6 citations), Li has helped shape the conversation around reliability in safety-critical AI. His work is essential reading for students and researchers building trustworthy autonomous systems.