PraxiMLP: A Threshold-based Framework for Efficient Three-Party MLP with Practical Security

TL;DR

PraxiMLP introduces a highly efficient three-party machine learning framework that operates entirely within the arithmetic domain, significantly enhancing precision and speed for privacy-preserving neural network training.

Contribution

It proposes novel additive-to-multiplicative conversion protocols that eliminate cross-domain conversions, supporting floating-point numbers for improved efficiency and accuracy.

Findings

8 orders of magnitude precision improvement over basic protocols

5x faster model training in WAN environments

Effective handling of non-linear functions with native floating-point support

Abstract

Efficiency and communication cost remain critical bottlenecks for practical Privacy-Preserving Machine Learning (PPML). Most existing frameworks rely on fixed-point arithmetic for strong security, which introduces significant precision loss and requires expensive cross-domain conversions (e.g., Arithmetic-to-Boolean) for non-linear operations. To address this, we propose PraxiMLP, a highly efficient three-party MLP framework grounded in practical security. The core of our work is a pair of novel additive-to-multiplicative conversion protocols that operate entirely within the arithmetic domain, thus avoiding expensive cross-domain conversions. By natively supporting loating-point numbers, PraxiMLP precisely handles non-linear functions, dramatically improving both efficiency and precision. Experimental results confirm that, compared to mainstream PPML frameworks, PraxiMLP delivers an average 8 orders of magnitude precision improvement on basic protocols and a 5x average model training speedup in a WAN environment.