• Scholar
  • Twitter
  • Instagram

CMU Safe AI Lab

Research laboratory at Carnegie Mellon University

The Safe AI Lab at CMU aims to develop reliable, explainable, verifiable, and good-for-all artificial intelligent learning methods for consequential applications in the face of the uncertain, dynamic, time-varying, multiple agents, and possibly human-involved environment by bridging fundamental learning theories and engineering technologies. The lab is recognized nationally and internationally for its research on autonomous vehicles and trustworthy AI. The lab has worked with many industrial partners, including Google Brain, Amazon, Ford, Uber, IBM, Adobe, Bosch, Toyota, and Rolls-Royce.

Positions are available for postdocs, master students, and interns. If you are interested in applying for a position, please fill out this form. It requires your CV and a 3-slide PPT describing your previous experience. For master and interns, we prefer students who have good engineering experience. For PhD application, we prefer students with publication experience. If you already have publication related to our research directions and a good GPA, you can send a reminder emails to zhaolab.jobs@gmail.com, by attaching one of your best papers and your CV. DO NOT repeatedly send me emails or send them to my other emails. It may make your email blocked.

We usually recruit 2 to 3 new Ph.D. students each year and have openings for postdocs targeting on faculty positions in top 20 universities. We recruite people at multiple departments of CMU, including Mechanical Engineering Department, Computer Science Department, and the Robotic Institute.

Recent News

  • [09/20/22] I am the keynote speaker for the 1st International Workshop on Safe Reinforcement Learning Theory and its Applications.
  • [08/31/22] I served as the co-chair for the Global Young Scientist Forum - AI and Intelligent Safety.
  • [08/30/22] Congrats Wenhao Ding and Jiacheng Zhu for winning the prestigious Qualcomm Innovation Fellowships (North America).
  • [07/26/22] I received the George Tallman Ladd Research Award. The award is made to a faculty member within the College of Engineering at CMU in recognition of outstanding research and professional accomplishments and potential.
  • [07/22/22] Our paper about robustness in safe RL won the best paper runner-up in the SL4AD Workshop at ICML 2022!.
  • [05/02/22] I joined the Robotic Team at Google Brain as a Visiting Researcher. I am still a full time professor at CMU. I chose to work at Google because I want to achieve the real-world impact at Google Scale and meet interesting souls.
  • [01/22/22] I was selected as one of the MIT Technical Review 35 Innovators under 35 China 2021.
  • [11/03/21] Our autonomous delivery robots were featured on CMU Engineering. We gave a demo to a group of legislators including four senators.
  • [10/19/21] I got the Ford University Collaboration Research Award
  • [09/30/21] Our new course Trustworthy AI Autonomy was featured on the front page of CMU's weekly news digest - the Piper.
  • [08/15/21] We will work on this very exciting five-year 10M AI-enabled Brain-Machine Interface project.
  • [05/21/21] I got the Carnegie-Bosch Award
  • [04/22/21] We got the award from Rolls-Royce to work on cybersecurity of real-time systems.
  • [03/17/21] I got the National Science Foundation CAREER Award.
  • [02/01/21] We are organizing the "Security and Safety in Machine Learning Systems" workshop in ICLR 2021. Please submit your papers here and win the best paper award!
  • [12/01/20] I got the Struminger Teaching Award.

Selected Recent Publications

Evaluation and Certification

M. Arief, Z. Huang, G. Kumar, Y. Bai, S. He, W. Ding, H. Lam, D. Zhao, ''Deep Probabilistic Accelerated Evaluation: A Robust Certifiable Rare-Event Simulation Methodology for Black-Box Safety-Critical Systems,'' Proceedings of the 24th International Conference on Artificial Intelligence and Statistics (AISTATS) 2021, San Diego, USA, 2021. | Bib

								@article{arief2020deep,
								title={Deep Probabilistic Accelerated Evaluation: A Robust Certifiable Rare-Event Simulation Methodology for Black-Box Safety-Critical Systems},
								author={Arief, Mansur and Huang, Zhiyuan and Kumar, Guru Koushik Senthil and Bai, Yuanlu and He, Shengyi and Ding, Wenhao and Lam, Henry and Zhao, Ding},
								journal={arXiv preprint arXiv:2006.15722},
								year={2021}
								}
								
F Wu, L Li, Z Huang, Y Vorobeychik, D Zhao, B Li, ''CROP: Certifying Robust Policies for Reinforcement Learning through Functional Smoothing,'' The International Conference on Learning Representations (ICLR), 2022. |

Safe Reinforcment Learning

Z Liu, Z Cen, V Isenbaev, W Liu, ZS Wu, B Li, D Zhao, ''Constrained variational policy optimization for safe reinforcement learning,'' Thirty-ninth International Conference on Machine Learning (ICML), 2022. |
B. Chen, Z. Liu, J. Zhu, M. Xu, W. Ding, D. Zhao, ''Context-Aware Safe Reinforcement Learning for Non-Stationary Environments,'' 2021 International Conference on Robotics and Automation (ICRA), 2021. | Bib

									@article{chen2021context,
				  title={Context-Aware Safe Reinforcement Learning for Non-Stationary Environments},
				  author={Chen, Baiming and Liu, Zuxin and Zhu, Jiacheng and Xu, Mengdi and Ding, Wenhao and Zhao, Ding},
				  journal={2021 International Conference on Robotics and Automation (ICRA)},
				  year={2021}
				}
								

Generalizable Reinforcment Learning

M. Xu, W. Ding, J. Zhu, Z. Liu, B. Chen, D. Zhao, ''Task-Agnostic Online Reinforcement Learning with an Infinite Mixture of Gaussian Processes,'' Advances in Neural Information Processing Systems (NeurIPS), 2020. | Bib

								@inproceedings{NEURIPS2020_47951a40,
								author = {Xu, Mengdi and Ding, Wenhao and Zhu, Jiacheng and LIU, ZUXIN and Chen, Baiming and Zhao, Ding},
								booktitle = {Advances in Neural Information Processing Systems (NeurIPS)},
								editor = {H. Larochelle and M. Ranzato and R. Hadsell and M. F. Balcan and H. Lin},
								pages = {6429--6440},
								publisher = {Curran Associates, Inc.},
								title = {Task-Agnostic Online Reinforcement Learning with an Infinite Mixture of Gaussian Processes},
								url = {https://proceedings.neurips.cc/paper/2020/file/47951a40efc0d2f7da8ff1ecbfde80f4-Paper.pdf},
								volume = {33},
								year = {2020}
								}
								
M. Xu, Y. Shen, S. Zhang, Y. Lu, D. Zhao, J, Tenenbaum, C. Gan, ''Prompting Decision Transformer for Few-shot Policy Generalization,'' Thirty-ninth International Conference on Machine Learning (ICML), 2022. |
W. Ding, H. Lin, B. Li, D. Zhao, ''Generalizing Goal-Conditioned Reinforcement Learning with Variational Causal Reasoning,'' Advances in Neural Information Processing Systems (NeurIPS), 2022. |

Survey and Benchmark

M Xu, Z Liu, P Huang, W Ding, Z Cen, B Li, D Zhao, ''Trustworthy Reinforcement Learning Against Intrinsic Vulnerabilities: Robustness, Safety, and Generalizability,'' arXiv preprint arXiv:2209.08025, 2022. |
W Ding, C Xu, H Lin, B Li, D Zhao, ''A Survey on Safety-critical Scenario Generation from Methodological Perspective,'' arXiv preprint arXiv:2202.02215, 2022. |
C Xu, W Ding, W Lyu, Z Liu, S Wang, Y He, H Hu, D Zhao, B Li, ''SafeBench: A Benchmarking Platform for Safety Evaluation of Autonomous Vehicles,'' Advances in Neural Information Processing Systems (NeurIPS), arXiv preprint arXiv:2206.09682, 2022. |
C Xie, Z Cao, Y Long, D Yang, D Zhao, B Li, ''Privacy of Autonomous Vehicles: Risks, Protection Methods, and Future Directions,'' arXiv preprint arXiv:2209.04022, 2022. |

Featured Research

Certifiable Accelerated Evaluation

Rigorous evaluation of intelligent physical systems against long-tail failures

  • More

Generating Safety-Critical Digital Twins

Use unsupervised learning, stochastic processing, and generative models to comprehend and generate naturalistic safety-critical environments

  • More

Trustworthy AI Autonomy

Design trustworthy intelligent autonomy for real world applications.

  • More

Autonomous Vehicle Platform

An open platform that facilitates development of Connected and Self-Driving Vehicles

  • More

Autonomous Mobile Robot Platform

Applications: multi-robot coordination in warehouse environment; food delivery robot on sidewalks.

  • More

Comprehend Human Being for Better Auto-Driving

Made by the people and for the people - drive harmoniously with human being

  • More

Affiliate Links

Menu

  • Homepage
  • People
  • Projects
  • Publications
  • Media
  • Teaching
  • Calendar

Contact

The best form of contact, generally, is via email.

  • dingzhao@cmu.edu
  • 412-268-3348
  • 5000 Forbes Avenue
    Pittsburgh, PA 15213

© Ding Zhao. All rights reserved. Design: HTML5 UP.