The paper investigates how temperature affects analog deep neural network (DNN) inference, focusing on both stochastic and systematic non‑idealities in analog hardware. Experiments show that temperature‑induced performance loss is mainly driven by systematic errors rather than random noise. The study evaluates various mitigation techniques, finding that noise‑aware training and temperature‑aware calibration—especially hardware‑in‑the‑loop training—best preserve inference accuracy across different thermal conditions.
By Niklas Summ, Xiao Wang, Hendrik Borras, Bernhard Klein, Holger Fr\"oning
arXiv:2606. 02781v1 Announce Type: cross Abstract: Deep neural networks (DNNs) have achieved state-of-the-art performance across diverse domains.
By Sohan Salahuddin Mugdho, Md. Shahedul Hasan, Brahmdutta Dixit, Yang Lv, Jian-Ping Wang, Cheng Wang
arXiv:2608. 10506v1 Announce Type: cross Abstract: Accurate pre-deployment estimation of CNN inference cost--energy, latency, and peak memory--is increasingly critical as models are deployed on resource-constrained GPU platforms.
By Linh Nguyen, Zhixin Pan
arXiv:2606. 23001v1 Announce Type: cross Abstract: On-device LLM inference is increasingly attractive for privacy-preserving, reliable, and cost-effective deployment, yet its energy and thermal costs remain a critical bottleneck.
By Bohua Zou, Nian Liu, Binqi Sun, Matteo Mascherin, Debayan Roy, Yutao Liu, Yu Peng, Ning Jia, Haibo Chen
Accurate pre-deployment estimation of CNN inference cost--energy, latency, and peak memory--is increasingly critical as models are deployed on resource-constrained GPU platforms. Existing approaches rely on FLOPs, latency measurements, or single-device profiling as energy proxies, overlooking the non-linear interactions between architectural design and hardware load.
The paper introduces SMART, a lightweight recalibration technique that adjusts logits based on the margin between the top two logits, called the logit gap. It uses a soft-binned Expected Calibration Error objective to balance bias and variance, enabling stable updates even with limited calibration data. Experiments across various datasets and architectures show SMART achieves state‑of‑the‑art calibration with fewer parameters than existing methods.
By Haolan Guo, Linwei Tao, Haoyang Luo, Minjing Dong, Chang Xu
arXiv:2607. 18101v1 Announce Type: new Abstract: On-device model adaptation is essential to enable lifelong personalization on resource-constrained hardware, but compute, power, and memory limitations of such devices make end-to-end backpropagation impractical for modern deep neural networks.
By Mateusz Piechocki, Alessandro Capotondi, Marek Kraft
The paper demonstrates that undervolting GPUs during CNN training introduces stochastic faults that act as implicit regularization, improving adversarial robustness while reducing power consumption. Experiments on LeNet, VGG-6, and MobileNetV3 trained on MNIST and CIFAR-10 show that undervolted models consistently outperform nominal-voltage models in both standard and adversarial training regimes. The approach offers a hardware-level defense that requires no algorithmic changes and yields significant energy savings due to the quadratic relationship between dynamic power and supply voltage.
By Behnam Omidi, Ahmad Tahmasivand, Husam Alsyouri, Saba Al-Sayouri, Chongzhou Fang, Ihsen Alouani, Khaled N. Khasawneh
arXiv:2607. 05933v1 Announce Type: cross Abstract: Dynamic Voltage Frequency Scaling (DVFS) on resource-constrained embedded GPU platforms is essential for energy-efficient small language model (SLM) fine-tuning, as privacy- and personalization-driven adaptation increasingly requires local execution and involves repeated forward-backward optimization over many mini-batches, making it substantially more time- and energy-intensive than single-pass inference.
By Jurn-Gyu Park, Sanzhar Zholdybayev, Aidar Amangeldi, Ademi Zhanuzakova
The paper introduces RAMP, a method for robust adaptive mixed‑precision quantization of vision models on edge CPUs. It evaluates 13 sensitivity metrics across four neural networks, finding that Jensen‑Shannon Divergence consistently identifies layers that can be safely quantized. Using K‑Means clustering on these metrics, RAMP achieves near‑lossless accuracy with an average 1.81× speed‑up, while cautioning against excluding low‑speed‑up layers that can fragment the computational graph.
By David Poblaci\'on-Criado, Dario Garcia-Gasulla, Eduardo Quinones
arXiv:2608. 02700v1 Announce Type: cross Abstract: Analog compute-in-memory (CIM) enables energy-efficient neural network inference, but device variation and read noise can severely degrade low-bit quantized models.
By Yizhe Chen, Wenshuai Yao, Saiya Wang, Yuannuo Feng, Wenbo Qi, Kechao Tang, Ngai Wong, Wenyong Zhou, Wang Kang
arXiv:2511. 15503v3 Announce Type: replace-cross Abstract: High-performance Host processors can integrate Processing-In-Memory (PIM) devices, which can accelerate memory-intensive kernels of Machine Learning (ML) models, including Large Language Models (LLMs), by leveraging the large memory bandwidth available at PIM cores.
By Peiming Yang, Sankeerth Durvasula, Ivan Fernandez, Mohammad Sadrosadati, Onur Mutlu, Gennady Pekhimenko, Christina Giannoula