Zitao Chen

Ayo~ I am a Ph.D. student in the ECE department at University of British Columbia (UBC), where I'm working in the Dependable Systems Lab with Prof. Karthik Pattabiraman. I received my M.A.Sc degree from UBC in 2020 and my bachelor degree from China University of Geosciences (Wuhan) in 2018. I am a UBC Public Scholar.

My research interest is on trustworthy machine learning (e.g., privacy, robustness, dependability).

Email: <first name><last initial>@ece.ubc.ca

Publications [Google Scholar]

Overconfidence is a Dangerous Thing: Mitigating Membership Inference Attacks by Enforcing Less Confident Prediction [paper] [code]
Zitao Chen, Karthik Pattabiraman
[NDSS'24] The ISOC Network and Distributed Systems Security Conference (Acceptance rate: 15%)
Artifact Available, Functional and Reproduced

Jujutsu: A Two-stage Defense against Adversarial Patch Attacks on Deep Neural Networks [paper] [slide] [code]
Zitao Chen, Pritam Dash, Karthik Pattabiraman
[AisaCCS'23] Proceedings of the 18th ACM ASIA Conference on Computer and Communications Security (Acceptance rate: 16%)

A Low-cost Fault Corrector for Deep Neural Networks through Range Restriction [paper] [slide] [code]
Zitao Chen, Guanpeng Li, Karthik Pattabiraman
[DSN'21] The 51st IEEE/IFIP International Conference on Dependable Systems and Networks (Acceptance rate: 16.3%)
Best Paper Award Runner-Up (2 out of 295 submissions)
Adopted by Intel OpenVINO [details]

PID-Piper: Recovering Robotic Vehicles from Physical Attacks [paper] [slide] [code]
Pritam Dash, Guanpeng Li, Zitao Chen, Mehdi Karimibiuki, and Karthik Pattabiraman
[DSN'21] The 51st IEEE/IFIP International Conference on Dependable Systems and Networks (Acceptance rate: 16.3%)
Best Paper Award (1 out of 295 submissions)

TensorFI: A Flexible Fault Injection Framework for TensorFlow Applications [paper] [slide] [code]
Zitao Chen*, Niranjhana Narayanan*, Bo Fang, Guanpeng Li, Karthik Pattabiraman, Nathan DeBardeleben
[ISSRE'20] The 31st International Symposium on Software Reliability Engineering (Acceptance rate: 25.7%)

BinFI: An Efficient Fault Injector for Safety-Critical Machine Learning Systems [paper] [slide] [code]
Zitao Chen, Guanpeng Li, Karthik Pattabiraman, Nathan DeBardeleben
[SC'19] International Conference for High Performance Computing, Networking, Storage, and Analysi (Acceptance rate: 20.9%)
Finalist for the SC reproducibility challenge (one of three papers)

Fault Injection for TensorFlow Applications [paper] [code]
Niranjhana Narayanan, Zitao Chen, Bo Fang, Guanpeng Li, Karthik Pattabiraman, Nathan DeBardeleben
[TDSC] IEEE Transactions on Dependable and Secure Computing

Awards
  • Brandwajn Graduate Award in Electrical and Computer Engineering
  • UBC Public Scholar
  • DSN Best Paper Award
  • DSN Best Paper Award Runner-up
  • UBC Four Year Fellowship (4YF)