ThRIve: Thermally Robust CNN Inference via Low-Rank Adaptation in Heterogeneous PIM Architectures

• URL: http://arxiv.org/abs/2607.17091v1
• Date: Sun, 19 Jul 2026 06:03:25 GMT
• Title: ThRIve: Thermally Robust CNN Inference via Low-Rank Adaptation in Heterogeneous PIM Architectures
• Abstract summary: Processing-In-Memory (PIM) has emerged as a promising technology for accelerating machine learning (ML) workloads.<n>PIM devices suffer from non-idealities such as thermal noise.<n>This noise alters the stored values in the memory cells which correspond to actual model weights, compromising the inference accuracy.<n>We introduce ThRIve, a noise-aware training methodology that leverages low-rank adaptation to enable thermally robust inference.
• Score: 13.754183593794131
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Processing-In-Memory (PIM) has emerged as a promising technology for accelerating machine learning (ML) workloads. Specifically, non-volatile memory-based PIM architectures have enabled effective ML acceleration due to their ability to perform energy-efficient matrix-vector multiplication operations. However, these devices suffer from non-idealities such as thermal noise. This noise alters the stored values in the memory cells which correspond to actual model weights, compromising the inference accuracy. In this work, we introduce ThRIve, a noise-aware training methodology that leverages low-rank adaptation to enable thermally robust inference on heterogeneous PIM architectures. ThRIve selectively stores these low-rank noise-aware parameters on a hardware that is less susceptible to thermal noise, enabling robustness against temperature-induced noise variations. ThRIve mitigates the effects of thermal-noise and prevent the drop in inference accuracy across the entire operating temperature range. Experimental results demonstrate that ThRIve-enabled architectures maintain consistent inference accuracy, with the mean accuracy staying within 2% of the ideal (i.e., noise-free) accuracy, and the variation in accuracy across the entire operating temperature range remaining within 2% of the mean. The proposed methodology achieves accuracy and robustness comparable to thermally-resilient Static Random-Access Memory (SRAM)-based PIM systems, while delivering up to 5.4x reduction in energy-delay product (EDP) during CNN model inferencing.

Related papers

• Beyond Noise: Understanding and Overcoming Temperature Effects in Analog DNN Inference [4.663767304979158]
  analog accelerators are susceptible to noise and non-idealities arising from physical component variations.<n>We show that temperature-induced degradation is driven primarily by systematic non-idealities rather than noise alone.<n>Noise-aware training improves robustness, while hardware-in-the-loop training and temperature calibration provide the strongest accuracy retention across varying thermal conditions.
  arXiv  Detail & Related papers  (2026-09-14T13:17:13Z)
• Scalable Lindblad Noise Learning via Stochastic Tensor-Network Simulation [48.887627688666974]
  Learning dissipation rates in large-scale open quantum systems is a major obstacle for near-term quantum technologies.<n>Here, we propose a scalable noise-learning framework for Lindblad dissipation rates that combines a simulation method, the Jump Method (TJM)<n>We demonstrate the approach on two noise models in the Ising model, in which independent dissipation rates are learned for each site up to $N_mathrmsite=16$.<n>This combination of scalable numerics and rigorous theoretical guarantees positions TJM-based noise learning as a practical foundation for characterizing dissipation in large quantum devices.
  arXiv  Detail & Related papers  (2026-08-25T15:07:11Z)
• Kinetics of sliding-window quantum error correction [39.34750401159749]
  We study the kinetics of sliding window decoding (SWD), an implementation of real-time decoding.<n>SWD acts on temporally local windows of noisy syndrome information and commits to corrections irreversibly at a nonzero rate.<n>We propose an effective description of SWD in terms of a kinetic process, where $mathbbZ$-charged point particles undergo parity-conserving reaction and diffusion.
  arXiv  Detail & Related papers  (2026-08-10T18:00:26Z)
• TIDY: Thermal Infrared Image Denoising via Wavelet Domain Entropy and Directional Stripe Index [7.740295165323681]
  Thermal infrared (TIR) imaging has been a popular choice for field robotics due to its robust perception capability under low light visual degradation.<n>It suffers from severe and fixed-pattern noise that breaks downstream estimation.<n>We propose TIDY, a lightweight wavelet-domain denoiser trained on real clean-noisy TIR data.
  arXiv  Detail & Related papers  (2026-06-18T05:42:50Z)
• Device-Agnostic Microwave Noise Metrology for Nonlinear Cryogenic Quantum Devices [33.608063888064514]
  Microwave devices capable of near-quantum-limited signal processing are essential components in the toolbox of solid-state quantum technologies.<n>The manipulation and readout of single-photon microwave signals through amplifiers, mixers, isolators, etc. must fulfill strict requirements in terms of signal integrity to ensure reliable operation.<n>This poses challenges to their characterization, since all relevant figures of merit must be expressed at the reference planes of their ports.
  arXiv  Detail & Related papers  (2026-05-27T17:57:07Z)
• Physics-Informed Machine Learning for Pouch Cell Temperature Estimation [5.435523531334183]
  This paper presents a physics-informed machine learning (PIML) framework for the efficient estimation of steady-state temperature profiles.<n>The PIML approach integrates the governing heat transfer equations directly into the neural network's loss function, enabling high-fidelity predictions.<n>Results demonstrate that the PIML model converges more rapidly and achieves markedly higher accuracy, with a 49.1% reduction in mean squared error over the data-driven model.
  arXiv  Detail & Related papers  (2026-04-16T02:58:10Z)
• Optimal multiparameter quantum estimation in accelerating Unruh-DeWitt detectors [0.09236074230806578]
  We investigate the ultimate precision limits for relativistic quantum thermometry in a bipartite system of uniformly accelerated detectors.<n>We show that dissipative noise results in more significant precision loss than purely dephasing mechanisms.
  arXiv  Detail & Related papers  (2026-03-08T01:35:45Z)
• Evaluating Temperature Scaling Calibration Effectiveness for CNNs under Varying Noise Levels in Brain Tumour Detection [0.18472148461613158]
  We develop a custom CNN and train it on a merged brain MRI dataset.<n>To simulate real-world uncertainty, five types of image noise are introduced: Gaussian, Poisson, Salt & Pepper, Speckle, and Uniform.<n>Results demonstrate that TS significantly reduces ECE and NLL under all noise conditions without degrading classification accuracy.
  arXiv  Detail & Related papers  (2025-09-29T15:46:23Z)
• Dataset Distillation with Neural Characteristic Function: A Minmax Perspective [39.77640775591437]
  We reformulate dataset distillation as a minmax optimization problem and introduce Neural Characteristic Function Discrepancy (NCFD)<n>NCFD is a comprehensive and theoretically grounded metric for measuring distributional differences.<n>Our method achieves significant performance gains over state-of-the-art methods on both low- and high-resolution datasets.
  arXiv  Detail & Related papers  (2025-02-28T02:14:55Z)
• PatchRefiner V2: Fast and Lightweight Real-Domain High-Resolution Metric Depth Estimation [38.71875790942604]
  PRV2 outperforms state-of-the-art depth estimation methods on UnrealStereo4K in both accuracy and speed.<n>It also shows improved depth boundary delineation on real-world datasets like CityScape, ScanNet++, and KITTI.
  arXiv  Detail & Related papers  (2025-01-02T07:41:27Z)
• FM2S: Towards Spatially-Correlated Noise Modeling in Zero-Shot Fluorescence Microscopy Image Denoising [33.383511185170214]
  Fluorescence Micrograph to Self (FM2S) is a zero-shot denoiser that achieves efficient Fluorescence Micrograph to Self (FM2S) denoising through three key innovations.<n>Experiments across FMI datasets demonstrate FM2S's superiority: It outperforms CVF-SID by 1.4dB PSNR on average while requiring 0.1% parameters of AP-BSN.
  arXiv  Detail & Related papers  (2024-12-13T10:45:25Z)
• Flow Matching for Atmospheric Retrieval of Exoplanets: Where Reliability meets Adaptive Noise Levels [38.84835238599221]
  Flow matching posterior estimation (FMPE) is a new machine learning approach to atmospheric retrieval. FMPE trains about 3 times faster than neural posterior estimation (NPE) and yields higher IS efficiencies. IS successfully corrects inaccurate ML results, identifies model failures via low efficiencies, and provides accurate estimates of the Bayesian evidence.
  arXiv  Detail & Related papers  (2024-10-28T19:28:07Z)
• Negative Pre-aware for Noisy Cross-modal Matching [46.5591267410225]
  Cross-modal noise-robust learning is a challenging task since noisy correspondence is hard to recognize and rectify. We present a novel Negative Pre-aware Cross-modal matching solution for large visual-language model fine-tuning on noisy downstream tasks.
  arXiv  Detail & Related papers  (2023-12-10T05:52:36Z)
• Improve Noise Tolerance of Robust Loss via Noise-Awareness [60.34670515595074]
  We propose a meta-learning method which is capable of adaptively learning a hyper parameter prediction function, called Noise-Aware-Robust-Loss-Adjuster (NARL-Adjuster for brevity) Four SOTA robust loss functions are attempted to be integrated with our algorithm, and comprehensive experiments substantiate the general availability and effectiveness of the proposed method in both its noise tolerance and performance.
  arXiv  Detail & Related papers  (2023-01-18T04:54:58Z)
• Dynamic Precision Analog Computing for Neural Networks [0.7701333337093469]
  We propose extending analog computing architectures to support varying levels of precision by repeating operations and averaging the result. We find that employing dynamic precision reduces energy consumption by up to 89% for computer vision models such as Resnet50 and by 24% for natural language processing models such as BERT.
  arXiv  Detail & Related papers  (2021-02-12T06:56:56Z)

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.