Guy Amit (Ben-Gurion University), Moshe Levy (Ben-Gurion University), Yisroel Mirsky (Ben-Gurion University)

Deep neural networks are normally executed in the forward direction. However, in this work, we identify a vulnerability that enables models to be trained in both directions and on different tasks. Adversaries can exploit this capability to hide rogue models within seemingly legitimate models. In addition, in this work we show that neural networks can be taught to systematically memorize and retrieve specific samples from datasets. Together, these findings expose a novel method in which adversaries can exfiltrate datasets from protected learning environments under the guise of legitimate models.

We focus on the data exfiltration attack and show that modern architectures can be used to secretly exfiltrate tens of thousands of samples with high fidelity, high enough to compromise data privacy and even train new models. Moreover, to mitigate this threat we propose a novel approach for detecting infected models.

View More Papers

Don't Interrupt Me – A Large-Scale Study of On-Device...

Marian Harbach (Google), Igor Bilogrevic (Google), Enrico Bacis (Google), Serena Chen (Google), Ravjit Uppal (Google), Andy Paicu (Google), Elias Klim (Google), Meggyn Watkins (Google), Balazs Engedy (Google)

Read More

Acoustic Keystroke Leakage on Smart Televisions

Tejas Kannan (University of Chicago), Synthia Qia Wang (University of Chicago), Max Sunog (University of Chicago), Abraham Bueno de Mesquita (University of Chicago Laboratory Schools), Nick Feamster (University of Chicago), Henry Hoffmann (University of Chicago)

Read More

OCPPStorm: A Comprehensive Fuzzing Tool for OCPP Implementations (Long)

Gaetano Coppoletta (University of Illinois Chicago), Rigel Gjomemo (Discovery Partners Institute, University of Illinois), Amanjot Kaur, Nima Valizadeh (Cardiff University), Venkat Venkatakrishnan (Discovery Partners Institute, University of Illinois), Omer Rana (Cardiff University)

Read More