Ditto: Quantization-aware Secure Inference of Transformers upon MPC

Due to the rising privacy concerns on sensitive client data and trained models like Transformers, secure multi-party computation (MPC) techniques are employed to enable secure inference despite attendant overhead. Existing works attempt to reduce the overhead using more MPC-friendly non-linear function approximations. However, the integration of quantization widely used in plaintext inference into the MPC domain remains unclear. To bridge this gap, we propose the framework named Ditto to enable more efficient quantization-aware secure Transformer inference. Concretely, we first incorporate an MPC-friendly quantization into Transformer inference and employ a quantization-aware distillation procedure to maintain the model utility. Then, we propose novel MPC primitives to support the type conversions that are essential in quantization and implement the quantization-aware MPC execution of secure quantized inference. This approach significantly decreases both computation and communication overhead, leading to improvements in overall efficiency. We conduct extensive experiments on Bert and GPT2 models to evaluate the performance of Ditto. The results demonstrate that Ditto is about $3.14\sim 4.40\times$ faster than MPCFormer (ICLR 2023) and $1.44\sim 2.35\times$ faster than the state-of-the-art work PUMA with negligible utility degradation.

SoK: Training Machine Learning Models over Multiple Sources with Privacy Preservation

Nowadays, gathering high-quality training data from multiple data controllers with privacy preservation is a key challenge to train high-quality machine learning models. The potential solutions could dramatically break the barriers among isolated data corpus, and consequently enlarge the range of data available for processing. To this end, both academia researchers and industrial vendors are recently strongly motivated to propose two main-stream folders of solutions: 1) Secure Multi-party Learning (MPL for short); and 2) Federated Learning (FL for short). These two solutions have their advantages and limitations when we evaluate them from privacy preservation, ways of communication, communication overhead, format of data, the accuracy of trained models, and application scenarios. Motivated to demonstrate the research progress and discuss the insights on the future directions, we thoroughly investigate these protocols and frameworks of both MPL and FL. At first, we define the problem of training machine learning models over multiple data sources with privacy-preserving (TMMPP for short). Then, we compare the recent studies of TMMPP from the aspects of the technical routes, parties supported, data partitioning, threat model, and supported machine learning models, to show the advantages and limitations. Next, we introduce the state-of-the-art platforms which support online training over multiple data sources. Finally, we discuss the potential directions to resolve the problem of TMMPP.