# DeepReShape: Redesigning Neural Networks for Efficient Private Inference

arXiv: https://arxiv.org/abs/2304.10593 · HTML: https://mlsanity.com/paper/2304.10593/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 20 Apr 2023.

## Built on by (most cited first, 8 of 8)

- [Verifiable evaluations of machine learning models using zkSNARKs](https://mlsanity.com/paper/2402.02675/) (2024)
- [ReLU's Revival: On the Entropic Overload in Normalization-Free Large Language Models](https://mlsanity.com/paper/2410.09637/) (2024)
- [Approximating ReLU on a Reduced Ring for Efficient MPC-based Private Inference](https://mlsanity.com/paper/2309.04875/) (2023)
- [CrypTorch: PyTorch-based Auto-tuning Compiler for Machine Learning with Multi-party Computation](https://mlsanity.com/paper/2511.19711/) (2025)
- [Regularized PolyKervNets: Optimizing Expressiveness and Efficiency for Private Inference in Deep Neural Networks](https://mlsanity.com/paper/2312.15229/) (2023)
- [PEEL the Layers and Find Yourself: Revisiting Inference-time Data Leakage for Residual Neural Networks](https://mlsanity.com/paper/2504.06410/) (2025)
- [DeepShare: Sharing ReLU Across Channels and Layers for Efficient Private Inference](https://mlsanity.com/paper/2512.17398/) (2025)
- [UFO: Unlocking Ultra-Efficient Quantized Private Inference with Protocol and Algorithm Co-Optimization](https://mlsanity.com/paper/2602.18758/) (2026)

## Abstract (arXiv)

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 strategically allocating channels to position the network's ReLUs in order of their criticality to network accuracy, 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 5.2× runtime improvement at iso-ReLU on CIFAR-100 and an 8.7× 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.