Robustness of on-device Models: Adversarial Attack to Deep Learning Models on Android Apps

Deep learning has shown its power in many applications. including object detection in images. natural-language understanding, and speech recogration. To make it more accessible to end users, many deep learning models are now embedded in mobile apps. Compared to offloading deep learning from smartphones to the cloud, performing machine learning on-device can help improve latency, connectivity, and power consumption. However, most deep learning models within Android apps can easily he obtained via mature reverse engineering, while the models' exposure may invite adversarial attacks. In this study, we propose a simple but effective approach to hacking deep learning models using adversarial attacks by identifying highly similar pre-trained models from TensorFlow Hub. All 10 real-world Android apps in the experiment are successfully attacked by our approach. Apart from the feasibility of the model attack, we also carry out an empirical study that investigates the characteristics of deep learning models used by hundreds of Android apps on Googly Play. The results show that many of them are similar to each other and widely use fine-tuning techniques to pre-trained models on the Internet.

Automated summary

The study shows that embedding deep learning models in mobile applications, such as Android apps, may be vulnerable to adversarial attacks. The experiment demonstrates that hackers can successfully attack real-world Android apps by identifying pre-trained models.