When Small Variations Become Big Failures: Reliability Challenges in Compute-in-Memory Neural Accelerators

• URL: http://arxiv.org/abs/2603.03491v1
• Date: Tue, 03 Mar 2026 20:08:25 GMT
• Title: When Small Variations Become Big Failures: Reliability Challenges in Compute-in-Memory Neural Accelerators
• Authors: Yifan Qin, Jiahao Zheng, Zheyu Yan, Wujie Wen, Xiaobo Sharon Hu, Yiyu Shi,
• Abstract summary: We show that even small device variations can lead to disproportionately high accuracy degradation and catastrophic failures in safety-critical inference workloads.<n>We introduce SWIM, a selective write-verify mechanism that strategically applies verification only where it is most impactful.<n>Finally, we explore a learning-centric solution that improves realistic worst-case performance by training networks with right-censored noise.
• Score: 10.520875052654771
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Compute-in-memory (CiM) architectures promise significant improvements in energy efficiency and throughput for deep neural network acceleration by alleviating the von Neumann bottleneck. However, their reliance on emerging non-volatile memory devices introduces device-level non-idealities-such as write variability, conductance drift, and stochastic noise-that fundamentally challenge reliability, predictability, and safety, especially in safety-critical applications. This talk examines the reliability limits of CiM-based neural accelerators and presents a series of techniques that bridge device physics, architecture, and learning algorithms to address these challenges. We first demonstrate that even small device variations can lead to disproportionately large accuracy degradation and catastrophic failures in safety-critical inference workloads, revealing a critical gap between average-case evaluations and worst-case behavior. Building on this insight, we introduce SWIM, a selective write-verify mechanism that strategically applies verification only where it is most impactful, significantly improving reliability while maintaining CiM's efficiency advantages. Finally, we explore a learning-centric solution that improves realistic worst-case performance by training neural networks with right-censored Gaussian noise, aligning training assumptions with hardware-induced variability and enabling robust deployment without excessive hardware overhead. Together, these works highlight the necessity of cross-layer co-design for CiM accelerators and provide a principled path toward dependable, efficient neural inference on emerging memory technologies-paving the way for their adoption in safety- and reliability-critical systems.

Related papers

• Blockchain-Enabled Routing for Zero-Trust Low-Altitude Intelligent Networks [77.17664010626726]
  We focus on the routing with multiple UAV clusters in low-altitude intelligent networks (LAINs)<n>To minimize the damage caused by potential threats, we present the zero-trust architecture with the software-defined perimeter and blockchain techniques.<n>We show that the proposed framework reduces the average E2E delay by 59% and improves the TSR by 29% on average compared to benchmarks.
  arXiv  Detail & Related papers  (2026-02-27T04:30:35Z)
• Towards Trustworthy Wi-Fi Sensing: Systematic Evaluation of Deep Learning Model Robustness to Adversarial Attacks [4.5835414225547195]
  We evaluate the robustness of CSI deep learning models under diverse threat models and varying degrees of attack realism.<n>Our experiments show that smaller models, while efficient and equally performant on clean data, are markedly less robust.<n>We confirm that physically realizable signal-space perturbations, designed to be feasible in real wireless channels, significantly reduce attack success.
  arXiv  Detail & Related papers  (2025-11-25T16:24:29Z)
• Communication-Efficient Federated Learning by Quantized Variance Reduction for Heterogeneous Wireless Edge Networks [55.467288506826755]
  Federated learning (FL) has been recognized as a viable solution for local-privacy-aware collaborative model training in wireless edge networks.<n>Most existing communication-efficient FL algorithms fail to reduce the significant inter-device variance.<n>We propose a novel communication-efficient FL algorithm, named FedQVR, which relies on a sophisticated variance-reduced scheme.
  arXiv  Detail & Related papers  (2025-01-20T04:26:21Z)
• Evaluating Single Event Upsets in Deep Neural Networks for Semantic Segmentation: an embedded system perspective [1.474723404975345]
  This paper delves into the robustness assessment in embedded Deep Neural Networks (DNNs)<n>By scrutinizing the layer-by-layer and bit-by-bit sensitivity of various encoder-decoder models to soft errors, this study thoroughly investigates the vulnerability of segmentation DNNs to SEUs.<n>We propose a set of practical lightweight error mitigation techniques with no memory or computational cost suitable for resource-constrained deployments.
  arXiv  Detail & Related papers  (2024-12-04T18:28:38Z)
• Towards Resource-Efficient Federated Learning in Industrial IoT for Multivariate Time Series Analysis [50.18156030818883]
  Anomaly and missing data constitute a thorny problem in industrial applications. Deep learning enabled anomaly detection has emerged as a critical direction. The data collected in edge devices contain user privacy.
  arXiv  Detail & Related papers  (2024-11-06T15:38:31Z)
• Digital Twin-Assisted Data-Driven Optimization for Reliable Edge Caching in Wireless Networks [60.54852710216738]
  We introduce a novel digital twin-assisted optimization framework, called D-REC, to ensure reliable caching in nextG wireless networks. By incorporating reliability modules into a constrained decision process, D-REC can adaptively adjust actions, rewards, and states to comply with advantageous constraints.
  arXiv  Detail & Related papers  (2024-06-29T02:40:28Z)
• Enhancing Dropout-based Bayesian Neural Networks with Multi-Exit on FPGA [20.629635991749808]
  This paper proposes an algorithm and hardware co-design framework that can generate field-programmable gate array (FPGA)-based accelerators for efficient BayesNNs. At the algorithm level, we propose novel multi-exit dropout-based BayesNNs with reduced computational and memory overheads. At the hardware level, this paper introduces a transformation framework that can generate FPGA-based accelerators for the proposed efficient BayesNNs.
  arXiv  Detail & Related papers  (2024-06-20T17:08:42Z)
• Center-Sensitive Kernel Optimization for Efficient On-Device Incremental Learning [88.78080749909665]
  Current on-device training methods just focus on efficient training without considering the catastrophic forgetting.<n>This paper proposes a simple but effective edge-friendly incremental learning framework.<n>Our method achieves average accuracy boost of 38.08% with even less memory and approximate computation.
  arXiv  Detail & Related papers  (2024-06-13T05:49:29Z)
• Enhancing Reliability of Neural Networks at the Edge: Inverted Normalization with Stochastic Affine Transformations [0.22499166814992438]
  We propose a method to inherently enhance the robustness and inference accuracy of BayNNs deployed in in-memory computing architectures. Empirical results show a graceful degradation in inference accuracy, with an improvement of up to $58.11%$.
  arXiv  Detail & Related papers  (2024-01-23T00:27:31Z)
• Scalable and Efficient Methods for Uncertainty Estimation and Reduction in Deep Learning [0.0]
  This paper explores scalable and efficient methods for uncertainty estimation and reduction in deep learning. We tackle the inherent uncertainties arising from out-of-distribution inputs and hardware non-idealities. Our approach encompasses problem-aware training algorithms, novel NN topologies, and hardware co-design solutions.
  arXiv  Detail & Related papers  (2024-01-13T19:30:34Z)
• Compute-in-Memory based Neural Network Accelerators for Safety-Critical Systems: Worst-Case Scenarios and Protections [8.813981342105151]
  We study the problem of pinpointing the worst-case performance of CiM accelerators affected by device variations. We propose a novel worst-case-aware training technique named A-TRICE that efficiently combines adversarial training and noise-injection training. Our experimental results demonstrate that A-TRICE improves the worst-case accuracy under device variations by up to 33%.
  arXiv  Detail & Related papers  (2023-12-11T05:56:00Z)
• Meta-Learning Priors for Safe Bayesian Optimization [72.8349503901712]
  We build on a meta-learning algorithm, F-PACOH, capable of providing reliable uncertainty quantification in settings of data scarcity. As core contribution, we develop a novel framework for choosing safety-compliant priors in a data-riven manner. On benchmark functions and a high-precision motion system, we demonstrate that our meta-learned priors accelerate the convergence of safe BO approaches.
  arXiv  Detail & Related papers  (2022-10-03T08:38:38Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.