Research Artifact of:
Research Artifact of:
CipherSteal: Stealing Input Data from TEE-Shielded Neural Networks with Ciphertext Side Channels
CipherSteal: Stealing Input Data from TEE-Shielded Neural Networks with Ciphertext Side Channels
Part 1. Experimental data
All our attacked neural networks and the collected side-channel traces are provided at:
Part 2. Code release
Our code is provided at: https://anonymous.4open.science/r/cipher-steal-D877
Examples of vulnerable modules in TVM exe, Glow exe, PyTorch, and TensorFlow are given in the README.
Part 3. More examples of recovery results of CipherSteal
Since thousands of examples are provided, we put them into a new page to accelerate the loading speed.
Click the "Examples" button below to see these examples.