Secure Floating-Point Training

ePrint Report: Secure Floating-Point Training

Deevashwer Rathee, Anwesh Bhattacharya, Divya Gupta, Rahul Sharma, Dawn Song

Secure 2-party computation (2PC) of floating-point arithmetic is improving in performance and recent work runs deep learning algorithms with it, while being as numerically precise as commonly used machine learning (ML) frameworks like PyTorch. We find that the existing 2PC libraries for floating-point support generic computations and lack specialized support for ML training. Hence, their latency and communication costs for compound operations (e.g., dot products) …

algorithms communication compound computation dawn song deep learning dot eprint report find frameworks high latency machine machine learning novel operations party performance point products protocols prove pytorch report song support training work