Rachel Luo

About

Rachel Luo is a leading researcher at the intersection of machine learning and high-stakes robotics, specializing in safety assurance and distribution shift detection. Her core contributions center on developing sample-efficient methods to ensure reliable robot performance when real-world conditions differ from training data. Luo’s most influential work, "Sample-Efficient Safety Assurances Using Conformal Prediction" (2022, 25 citations), introduces a pioneering framework that leverages conformal prediction to provide early warning systems for unsafe situations, enabling robots to operate with provable safety guarantees even with limited data. She has also advanced the field through her system-level analysis of out-of-distribution data in robotics (2022, 7 citations), offering a comprehensive view of how learned components fail under distribution shift. More recently, Luo has tackled the challenge of online shift detection via recency prediction (2024, 2 citations), addressing the critical need for real-time monitoring in streaming robotics applications. Her work bridges theoretical rigor with practical deployment, making her a key voice in trustworthy autonomous systems. With over 50 total citations, Luo’s research is shaping how we build robots that can safely navigate unpredictable environments.

Research Focus

Key Achievements

Top Papers

1. 1
   Sample-Efficient Safety Assurances Using Conformal Prediction

   25 citations · 2022
2. 2
   Sample-efficient safety assurances using conformal prediction

   10 citations · 2023
3. 3
   A System-Level View on Out-of-Distribution Data in Robotics

   7 citations · 2022
4. 4
   Sample-Efficient Safety Assurances using Conformal Prediction

   4 citations · 2021
5. 5
   Online Distribution Shift Detection via Recency Prediction

   2 citations · 2024
6. 6
   Online Distribution Shift Detection via Recency Prediction

   2 citations · 2022

Key Collaborators

Contact & Links