Chiao Hsieh

Papers

Total Citations

H-Index

About

Chiao Hsieh is a researcher working at the intersection of formal verification, autonomous systems, and machine learning safety — tackling one of the most pressing challenges in modern robotics: how do we rigorously guarantee the safety of systems that rely on imperfect perception? His work addresses the critical gap between theoretical verification and the messy realities of vision-based and neural perception models integrated into control systems. Hsieh's most influential contribution, "Verifying Controllers With Vision-Based Perception Using Safe Approximate Abstractions" (2022, 32 citations), introduces a practical framework for reasoning about safety even when full formal verification of perception models remains intractable. Building on this, his 2023 work on "Perception Contracts" (15 citations) develops a principled theory for bounding acceptable perception errors while preserving system-level safety invariants — a conceptually elegant advance for ML-enabled autonomous systems. His earlier work on Koord (2020, 17 citations) demonstrates his breadth, providing a domain-specific programming language that simplifies developing and verifying distributed multi-robot applications without requiring deep hardware expertise. Collectively, Hsieh's research equips engineers and theorists alike with tools to build provably safer autonomous systems, making him a notable emerging voice in the formal methods and robotics communities.