Reaching Data Confidentiality and Model Accountability on the CalTrain
This addresses security and accountability issues in multi-party collaborative learning for applications where data privacy and model trust are critical, representing a novel integration rather than an incremental improvement.
The paper tackles the vulnerability of distributed collaborative learning to data poisoning and backdoor attacks by introducing CALTRAIN, a system that uses a Trusted Execution Environment to achieve data confidentiality and model accountability, with evaluation showing it maintains prediction accuracy and accurately identifies malicious training data.
Distributed collaborative learning (DCL) paradigms enable building joint machine learning models from distrusting multi-party participants. Data confidentiality is guaranteed by retaining private training data on each participant's local infrastructure. However, this approach to achieving data confidentiality makes today's DCL designs fundamentally vulnerable to data poisoning and backdoor attacks. It also limits DCL's model accountability, which is key to backtracking the responsible "bad" training data instances/contributors. In this paper, we introduce CALTRAIN, a Trusted Execution Environment (TEE) based centralized multi-party collaborative learning system that simultaneously achieves data confidentiality and model accountability. CALTRAIN enforces isolated computation on centrally aggregated training data to guarantee data confidentiality. To support building accountable learning models, we securely maintain the links between training instances and their corresponding contributors. Our evaluation shows that the models generated from CALTRAIN can achieve the same prediction accuracy when compared to the models trained in non-protected environments. We also demonstrate that when malicious training participants tend to implant backdoors during model training, CALTRAIN can accurately and precisely discover the poisoned and mislabeled training data that lead to the runtime mispredictions.