The Safe Artificial Intelligence and Cyber-Physical Systems (Safe-AI-CPS) Lab is directed by Prof. Bhaskar Ramasubramanian. We are located in the Department of Engineering and Design at Western Washington University.
Our research objective is to reason about the behavior of autonomous cyber-physical systems (CPS) using techniques from machine learning, control, optimization, and game theory. It is becoming increasingly common for such CPS to share an environment with humans (e.g., self-driving and human-driven cars). In this setting, we are interested in developing solutions that: i) integrate feedback from multiple heterogeneous sources, ii) are resilient to actions of malicious/ dishonest participants, and iii) provide provable performance guarantees.
[August 2025] One paper accepted to the IEEE Transactions on Dependable and Secure Computing (IEEE TDSC).
[July 2025] One paper accepted to the IEEE Conference on Decision and Control (IEEE CDC) 2025.
[May 2025] EECE Sophomore John Lewis has been awarded a McCabe Summer Research Award. Congratulations, John!
[May 2025] Our paper titled `Small Models Struggle to Learn from Strong Reasoners' has been accepted to the Findings of the Annual Meeting of the Association for Computational Linguistics (ACL), 2025. This is joint work with Yuetai Li, Zhangchen Xu, Fengqing Jiang, Bill Yuchen Lin, Luyao Niu and Radha Poovendran at UW, and Xiang Yue at CMU.
[Jan 2025] Our paper titled `Modeling and Designing Non-Pharmaceutical Interventions in Epidemics: A Submodular Approach' has been accepted to the American Control Conference 2025. This is joint work with Shiyu Cheng and Andrew Clark at Washington University in St. Louis, and Luyao Niu and Radha Poovendran at UW.
[Nov 2024] Our paper titled `Modeling and Designing Non-Pharmaceutical Interventions in Epidemics: A Submodular Approach' has been accepted to the IEEE Control System Letters (L-CSS). This is joint work with Shiyu Cheng and Andrew Clark at Washington University in St. Louis, and Luyao Niu and Radha Poovendran at UW.
[Nov 2024] Our paper titled `CANTXSec: A Deterministic Intrusion Detection and Prevention System for CAN Bus Monitoring ECU Activations' has been accepted to the 2025 International Conference on Applied Cryptography and Network Security (ACNS). This is joint work with Denis Donadel, Alessandro Brighente, and Mauro Conti at the University of Padua, and Kavya Balasubramanian and Radha Poovendran at UW.
[Sep 2024] Our paper titled `CleanGen: Mitigating Backdoor Attacks for Generation Tasks in Large Language Models' has been accepted to the 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP). This is joint work with Yuetai Li, Zhangchen Xu, Fengqing Jiang, Luyao Niu, Dinuka Sahabandu, and Radha Poovendran at UW.
[Sep 2024] Our paper titled `Who is Responsible? Explaining Safety Violations in Multi-Agent Cyber Physical Systems' has been accepted to the IEEE International Conference on Assured Autonomy (ICAA), 2024. This is joint work with Luyao Niu, Dinuka Sahabandu, and Radha Poovendran at UW and Hongchao Zhang and Andrew Clark at WUSTL.
[Sep 2024] Our paper titled `Rapid Autonomy Transfer in Reinforcement Learning with a Single Pretrained Critic' has been accepted to the IEEE International Conference on Tools with Artificial Intelligence (ICTAI), 2024. This is joint work with Muhammad Faraz Karim, Yunjie Deng, Michail Alexiou, and Sukarno Mertoguno at GATech, and Luyao Niu, Dinuka Sahabandu, and Radha Poovendran at UW.
[Jul 2024] Our paper titled `Sampling and Quantization-Aware Control Barrier Functions for Safety-Critical Cyber-Physical Systems' has been accepted to the IEEE Conference on Decision and Control (CDC), 2024. This is joint work with Luyao Niu and Radha Poovendran at UW, and Andrew Clark at Washington Univ. in St. Louis.
[May 2024] Our paper titled `ArtPrompt: ASCII Art-based Jailbreak Attacks against Aligned LLMs' has been accepted to the Annual Meeting of the Association for Computational Linguistics (ACL), 2024. This is joint work with Fengqing Jiang, Zhangchen Xu, Luyao Niu, and Radha Poovendran at UW, Zhen Xiang at UIUC, and Bo Li at Univ. of Chicago/ UIUC. This work has received extensive media coverage.
[Mar 2024] Our paper titled `EDC: Effective and Efficient Dialog Comprehension for Dialog State Tracking' has been accepted to the 2024 Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL). This is joint work with Qifan Lu and Radha Poovendran at UW.
[Mar 2024] EECE Junior Jared Hechter has been awarded the WWU CSE Jarvis Memorial Summer Research Award. Congratulations, Jared!!
[Mar 2024] Our paper titled `ArtPrompt: ASCII Art-based Jailbreak Attacks against Aligned LLMs' has been accepted to the ICLR 2024 Workshop on Secure and Trustworthy LLMs (SET LLM). This is joint work with Fengqing Jiang, Zhangchen Xu, Luyao Niu, and Radha Poovendran at UW, Zhen Xiang at UIUC, and Bo Li at Univ. of Chicago/ UIUC.
[Jan 2024] Our paper titled `Risk-Aware Distributed Multi-Agent Reinforcement Learning' has been accepted to the American Control Conference (ACC) 2024. This is joint work with Abdullah Al Maruf at CalState LA, Andrew Clark at Washington Univ. in St. Louis, Luyao Niu and Radha Poovendran at UW.
[Jan 2024] Our paper titled `BadChain: Backdoor Chain-of-Thought Prompting for Large Language Models' has been accepted to the International Conference on Learning Representations (ICLR) 2024. This is joint work with Zheng Xiang and Zidi Xiong at UIUC, Bo Li at Univ. of Chicago/ UIUC, Fengqing Jiang and Radha Poovendran at UW.
[Dec 2023] Our paper titled `BadChain: Backdoor Chain-of-Thought Prompting for Large Language Models' was presented at the NeurIPS 2023 Workshop on Backdoors in Deep Learning- The Good, The Bad, and The Ugly. This is joint work with Zheng Xiang and Zidi Xiong at UIUC, Bo Li at Univ. of Chicago/ UIUC, Fengqing Jiang and Radha Poovendran at UW.
[Nov 2023] Our paper titled `CANLP: NLP-Based Intrusion Detection System for CAN' has been accepted to the ACM Symposium on Applied Computing. This is joint work with Denis Donadel, Alessandro Brighente, and Mauro Conti at the University of Padua, and Kavya Balasubramanian, Adithya Gowda Baragur, Dinuka Sahabandu, and Radha Poovendran at UW.
[Sep 2023] Our paper titled `FedGame: A Game-Theoretic Defense Against Backdoor Attacks in Federated Learning' has been accepted to the Conference on Neural Information Processing Systems (NeurIPS). This is joint work with Jinyuan Jia at Penn State, Zhuowen Yuan at UIUC, Bo Li at Univ. of Chicago/ UIUC, Arezoo Rajabi, Dinuka Sahabandu, Luyao Niu, and Radha Poovendran at UW.
[Aug 2023] Our paper titled `MDTD: A Multi-Domain Trojan Detector for Deep Neural Networks' has been accepted to the ACM Conference on Computer and Communications Security (CCS). This is joint work with Arezoo Rajabi, Surudhi Asokraj, Fengqing Jiang, Luyao Niu, Jim Ritcey, and Radha Poovendran at UW.
[May 2023] Dominic Danis and Ryan Blake present posters at WWU Scholars Week 2023. Both posters attracted an enthusiastic audience from the WWU Community!
[May 2023] Our paper titled `Learning Dissemination Strategies for External Sources in Opinion Dynamic Models with Cognitive Biases' has been accepted to the International Joint Conference on Artificial Intelligence (IJCAI). This is joint work with Abdullah Al Maruf, Luyao Niu, and Radha Poovendran at UW, and Andrew Clark at WUSTL.
[Mar 2023] Our paper titled `BARON: Base-Station Authentication through Core Network for Mobility Management in 5G Networks' has been accepted to the ACM International Conference on Security and Privacy for Wireless and Mobile Networks, 2023. This is joint work with Alessandro Lotto, Alessandro Brighente, and Mauro Conti at the University of Padua, and Vaibhav Singh and Radha Poovendran at UW.
[Mar 2023] Our paper titled `Robust Satisfaction of Metric Interval Temporal Logic Objectives in Adversarial Environments' has been accepted to MDPI Games Journal Special Issue on Game-Theoretic Analysis of Network Security and Privacy. The article has also been selected to feature on the Issue Cover Page! This is joint work with Luyao Niu and Radha Poovendran at UW, and Andrew Clark at WUSTL.
[Mar 2023] Our paper titled `Multi-Agent Reinforcement Learning with Prospect Theory' has been accepted to the SIAM Conference on Control and its Applications, 2023. Congratulations to WWU EECE Undergrads Dominic Danis, Parker Parmacek, and David Dunajsky for all the hard work!!
[Mar 2023] EECE Junior Tran Minh Khoi Le has been awarded the WWU CSE Jarvis Memorial Summer Research Award. Congratulations, Khoi!!
[Jan 2023] Our paper titled `Cognitive Bias-Aware Dissemination Strategies for Opinion Dynamics with External Information Sources' has been accepted to the International Conference on Autonomous Agents and Multi Agent Systems, 2023 as an Extended Abstract and Poster Presentation. This is joint work with Abdullah Al Maruf, Luyao Niu, and Radha Poovendran at UW, and Andrew Clark at WUSTL.
[Nov 2022] Our paper titled `LDL: A Defense for Label-Based Membership Inference Attacks' has been accepted to the 18th ACM ASIA Conference on Computer and Communications Security 2023. This is joint work with Arezoo Rajabi, Dinuka Sahabandu, Luyao Niu, and Radha Poovendran at UW.
[Nov 2022] Attended the NSF CPS PI Meeting 2022 in Washington, D.C., and presented a poster for my CRII grant.
[Jul 2022] Our paper titled `Privacy-Preserving Reinforcement Learning Beyond Expectation' has been accepted to the 61st IEEE Conference on Decision and Control. This is joint work with Arezoo Rajabi, Abdullah Al Maruf, and Radha Poovendran at UW.
[Jun 2022] Five students will be working in the lab during Summer 2022. Dominic and Tyler will continue their work from Winter/ Spring 2022. Welcome, Minh, David, and Parker!
[Apr 2022] Three students join the lab. Welcome, Dominic, Tyler, and Kris! Dominic and Tyler will be working on the NSF-supported CRII project; Kris is working to design and build a system for particulate matter detection.
[Jan 2022] Awarded a CRII grant from the NSF for a project titled Cognizant Learning for Autonomous Cyber-Physical Systems.
[Jan 2022] Our paper titled `Agent-Temporal Attention for Reward Redistribution in Episodic Multi-Agent Reinforcement Learning' has been accepted to the International Conference on Autonomous Agents and Multi-Agent Systems, 2022 as an Oral Presentation. This is joint work with Baicen Xiao and Radha Poovendran at UW.
[Dec 2021] Presented our paper titled `Reinforcement Learning Beyond Expectation' at the 60th Conference on Decision and Control (Virtual). This is joint work with Luyao Niu and Andrew Clark at WPI, and Radha Poovendran at UW.