Machine learning (ML) has established itself as a cornerstone for various critical applications ranging from autonomous driving to authentication systems. However, with this increasing adoption rate of machine learning models, multiple attacks have emerged. One class of such attacks is training time attack, whereby an adversary executes their attack before or during the machine learning model training. In this work, we propose a new training time attack against computer vision based machine learning models, namely model hijacking attack. The adversary aims to hijack a target model to execute a different task than its original one without the model owner noticing. Model hijacking can cause accountability and security risks since a hijacked model owner can be framed for having their model offering illegal or unethical services. Model hijacking attacks are launched in the same way as existing data poisoning attacks. However, one requirement of the model hijacking attack is to be stealthy, i.e., the data samples used to hijack the target model should look similar to the model's original training dataset. To this end, we propose two different model hijacking attacks, namely Chameleon and Adverse Chameleon, based on a novel encoder-decoder style ML model, namely the Camouflager. Our evaluation shows that both of our model hijacking attacks achieve a high attack success rate, with a negligible drop in model utility.

View More Papers

The Inconvenient Truths of Ground Truth for Binary Analysis

Jim Alves-Foss, Varsha Venugopal (University of Idaho)

Read More

Binary Search in Secure Computation

Marina Blanton (University at Buffalo (SUNY)), Chen Yuan (University at Buffalo (SUNY))

Read More

The Taming of the Stack: Isolating Stack Data from...

Kaiming Huang (Penn State University), Yongzhe Huang (Penn State University), Mathias Payer (EPFL), Zhiyun Qian (UC Riverside), Jack Sampson (Penn...

Read More

Local and Central Differential Privacy for Robustness and Privacy...

Mohammad Naseri (University College London), Jamie Hayes (DeepMind), Emiliano De Cristofaro (University College London & Alan Turing Institute)

Read More