DeepReShape: Redesigning Neural Networks for Efficient Private Inference

DeepReShape network redesigning pipeline


Prior work on Private Inference (PI)–inferences performed directly on encrypted input–has focused on minimizing a network’s ReLUs, which have been assumed to dominate PI latency rather than FLOPs. Recent work has shown that FLOPs for PI can no longer be ignored and incur high latency penalties. In this paper, we develop DeepReShape, a technique that optimizes neural network architectures under PI’s constraints, optimizing for both ReLUs and FLOPs for the first time. The key insight is that a strategic allocation of channels such that the network’s ReLUs are positioned in their criticality order simultaneously optimizes ReLU and FLOPs efficiency. DeepReShape automates network development with an efficient process, and we call generated networks HybReNets. We evaluate DeepReShape using standard PI benchmarks and demonstrate a 2.1% accuracy gain with a 5.2x runtime improvement at iso-ReLU on CIFAR-100 and an 8.7x runtime improvement at iso-accuracy on TinyImageNet. Furthermore, we investigate the significance of network selection in prior ReLU optimizations and shed light on the key network attributes for superior PI performance.

TMLR 2024
Nandan Kumar Jha
Nandan Kumar Jha
PhD student at NYU CCS

My research goal is to enable near-real-time inference on encrypted data by co-designing deep neural networks and cryptographic primitives.