# Training Modern Deep Neural Networks for Memory-Fault Robustness

arXiv: https://arxiv.org/abs/1911.10287 · HTML: https://mlsanity.com/paper/1911.10287/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 23 Nov 2019.

## Built on by (most cited first, 7 of 7)

- [FT-Transformer: Resilient and Reliable Transformer with End-to-End Fault Tolerant Attention](https://mlsanity.com/paper/2504.02211/) (2025)
- [Winograd Convolution: A Perspective from Fault Tolerance](https://mlsanity.com/paper/2202.08675/) (2022)
- [Fault-Tolerant Deep Learning: A Hierarchical Perspective](https://mlsanity.com/paper/2204.01942/) (2022)
- [MemSE: Fast MSE Prediction for Noisy Memristor-Based DNN Accelerators](https://mlsanity.com/paper/2205.01707/) (2022)
- [Exploring Winograd Convolution for Cost-effective Neural Network Fault Tolerance](https://mlsanity.com/paper/2308.08230/) (2023)
- [Optimizing the Energy Efficiency of Unreliable Memories for Quantized Kalman Filtering](https://mlsanity.com/paper/2109.01520/) (2021)
- [Long-Term and Short-Term Transistor Aging in Deep Neural Networks: Impact and Mitigation](https://mlsanity.com/paper/2606.04266/) (2026)

## Abstract (arXiv)

Because deep neural networks (DNNs) rely on a large number of parameters and computations, their implementation in energy-constrained systems is challenging. In this paper, we investigate the solution of reducing the supply voltage of the memories used in the system, which results in bit-cell faults. We explore the robustness of state-of-the-art DNN architectures towards such defects and propose a regularizer meant to mitigate their effects on accuracy. Our experiments clearly demonstrate the interest of operating the system in a faulty regime to save energy without reducing accuracy.