BEGIN:VCALENDAR BEGIN:VEVENT SUMMARY:Center of Autonomy and AI, Center for Cyber-Physical Systems and the Internet of Things, and Ming Hsieh Institute Seminar Series DESCRIPTION:Speaker: Haiying (Helen) Shen , Department of Computer Science, University of Virginia Talk Title: Black-box Adversarial Attacks for Deep Driving Maneuver Classification Models in Transportation CPS Series: Center for Cyber-Physical Systems and Internet of Things Abstract: In a connected autonomous vehicle (CAV) scenario, each vehicle utilizes an onboard deep neural network (DNN) model to understand its received time-series driving signals (e.g., speed, brake status) from its nearby vehicles, and then takes necessary actions to increase traffic safety and roadway efficiency. In the scenario, an attacker may launch an adversarial attack, in which the attacker adds unnoticeable perturbation to the actual driving signals to fool the DNN model inside a victim vehicle to output a misclassified class to cause traffic congestion and/or accidents. Such an attack must be generated in near real-time and the adversarial perturbation is not noticeable in the current traffic context. However, previously proposed adversarial attacks fail to meet these requirements. To handle these challenges, we propose black-box adversarial attacks for time-series DNN models in the CAV scenarios. By analyzing real driving datasets, we observe driving signal patterns and features. Then, based on our observations, we design offline perturbations, which are used as a starting point in the online perturbation determination to generate the attacks. Our extensive experimental studies using two real driving datasets show that our proposed adversarial attacks require much shorter generation time and less perturbation amount than existing adversarial attacks.\n \n Biography: Dr. Haiying (Helen) Shen is currently an Associate Professor in the Department of Computer Science at University of Virginia. She received the 2015 IEEE Technical Committee on Scalable Computing (TCSC) Mid-career Award, the 2010 Microsoft Faculty Fellowship Award, the 2015 IBM Faculty Award, the 2013 NSF CAREER Award, and the 2013 Sigma Xi Clemson Chapter Young Investigator Award. Her research interests include Cyber-physical systems, Cloud computing and datacenters, Machine learning, Big data, and Distributed systems. Dr. Shen has made substantial contributions to her field with over 330 publications in prestigious conferences and journals such as Sigcomm'21, CoNext'20, Infocom'2011-2019, and IEEE/ACM Transactions on Networking (TON). Her papers received George N. Saridis best transactions paper award2021, the best paper awards in CloudCom2016 and NAS2018, best paper runner-up award in ICCCN2015, best paper award nominees in ICPP2021, MASS2011 and CCGrid2009, and best-in-session-presentation award in INFOCOM2017. She currently advises five Ph.D. students. She is an Associate Editor for the IEEE/ACM Transactions on Networking (TON), IEEE Transactions on Mobile Computing (TMC), IEEE Networking Letters (NL). She is also a program committee member of many leading conferences, and the former program co-chair and general co-chair for a number of international conferences. Host: Pierluigi Nuzzo and Bhaskar Krishnamachari Webcast: https://usc.zoom.us/webinar/register/WN_p5OEJlPxQlakO4hqovuGEQ DTSTART:20211020T140000 LOCATION: Online URL;VALUE=URI: DTEND:20211020T150000 END:VEVENT END:VCALENDAR