Scale When Needed: Adaptive Neuron-level Mixed Precision Quantization Aware Training

• URL: http://arxiv.org/abs/2605.25054v1
• Date: Sun, 24 May 2026 13:02:42 GMT
• Title: Scale When Needed: Adaptive Neuron-level Mixed Precision Quantization Aware Training
• Abstract summary: Quantization-Aware Training (QAT) has emerged as a leading compression approach.<n>We propose Neuron-Level Mixed-Precision QAT (NMP-QAT)<n>NMP-QAT achieves superior compression-accuracy trade-offs over mixed-precision QAT baselines.
• Score: 0.06157382820537719
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Deploying deep neural networks on resource-constrained 6G edge devices demands aggressive compression with minimal accuracy loss. Quantization-Aware Training (QAT) has emerged as a leading compression approach; however, existing mixed-precision methods typically operate at coarse layer- or channel-level granularity. These methods often rely on heuristic or search-based bit-allocation strategies, which may overlook fine-grained variability at the neuron level. We propose Neuron-Level Mixed-Precision QAT (NMP-QAT), where each neuron independently learns its own discrete precision during training. Starting from low-bit precision, NMP-QAT expands bit-width only when training signals demand it, via differentiable surrogates and straight-through estimators, while preserving a fully discrete inference graph. This adaptability extends to both weights and activations, reducing memory movement. Evaluated on telecom and non-telecom datasets across MLP and tabular foundation model architectures, NMP-QAT achieves superior compression-accuracy trade-offs over mixed-precision QAT baselines, making it well-suited for Green AI deployments at the network edge.

Related papers

• Mixed Precision Training of Neural ODEs [1.3382837742547355]
  This paper presents a mixed precision training framework for neural ODEs.<n>It combines explicit ODE solvers with a custom backpropagation scheme.<n>It achieves approximately 50% memory reduction and up to 2x speedup while maintaining accuracy comparable to single-precision training.
  arXiv  Detail & Related papers  (2025-10-27T16:32:56Z)
• Progressive Element-wise Gradient Estimation for Neural Network Quantization [2.1413624861650358]
  Quantization-Aware Training (QAT) methods rely on the Straight-Through Estimator (STE) to address the non-differentiability of discretization functions.<n>We propose Progressive Element-wise Gradient Estimation (PEGE) to address discretization errors between continuous and quantized values.<n>PEGE consistently outperforms existing backpropagation methods and enables low-precision models to match or even outperform the accuracy of their full-precision counterparts.
  arXiv  Detail & Related papers  (2025-08-27T15:59:36Z)
• Quantization Aware Factorization for Deep Neural Network Compression [20.04951101799232]
  decomposition of convolutional and fully-connected layers is an effective way to reduce parameters and FLOP in neural networks. A conventional post-training quantization approach applied to networks with weights yields a drop in accuracy. This motivated us to develop an algorithm that finds decomposed approximation directly with quantized factors.
  arXiv  Detail & Related papers  (2023-08-08T21:38:02Z)
• Efficient Uncertainty Quantification and Reduction for Over-Parameterized Neural Networks [23.7125322065694]
  Uncertainty quantification (UQ) is important for reliability assessment and enhancement of machine learning models. We create statistically guaranteed schemes to principally emphcharacterize, and emphremove, the uncertainty of over- parameterized neural networks. In particular, our approach, based on what we call a procedural-noise-correcting (PNC) predictor, removes the procedural uncertainty by using only emphone auxiliary network that is trained on a suitably labeled dataset.
  arXiv  Detail & Related papers  (2023-06-09T05:15:53Z)
• CSQ: Growing Mixed-Precision Quantization Scheme with Bi-level Continuous Sparsification [51.81850995661478]
  Mixed-precision quantization has been widely applied on deep neural networks (DNNs) Previous attempts on bit-level regularization and pruning-based dynamic precision adjustment during training suffer from noisy gradients and unstable convergence. We propose Continuous Sparsification Quantization (CSQ), a bit-level training method to search for mixed-precision quantization schemes with improved stability.
  arXiv  Detail & Related papers  (2022-12-06T05:44:21Z)
• AMED: Automatic Mixed-Precision Quantization for Edge Devices [3.5223695602582614]
  Quantized neural networks are well known for reducing the latency, power consumption, and model size without significant harm to the performance. Mixed-precision quantization offers better utilization of customized hardware that supports arithmetic operations at different bitwidths.
  arXiv  Detail & Related papers  (2022-05-30T21:23:22Z)
• Mixed Precision Low-bit Quantization of Neural Network Language Models for Speech Recognition [67.95996816744251]
  State-of-the-art language models (LMs) represented by long-short term memory recurrent neural networks (LSTM-RNNs) and Transformers are becoming increasingly complex and expensive for practical applications. Current quantization methods are based on uniform precision and fail to account for the varying performance sensitivity at different parts of LMs to quantization errors. Novel mixed precision neural network LM quantization methods are proposed in this paper.
  arXiv  Detail & Related papers  (2021-11-29T12:24:02Z)
• Low-bit Quantization of Recurrent Neural Network Language Models Using Alternating Direction Methods of Multipliers [67.688697838109]
  This paper presents a novel method to train quantized RNNLMs from scratch using alternating direction methods of multipliers (ADMM) Experiments on two tasks suggest the proposed ADMM quantization achieved a model size compression factor of up to 31 times over the full precision baseline RNNLMs.
  arXiv  Detail & Related papers  (2021-11-29T09:30:06Z)
• Neural Architecture Search for Efficient Uncalibrated Deep Photometric Stereo [105.05232615226602]
  We leverage differentiable neural architecture search (NAS) strategy to find uncalibrated PS architecture automatically. Experiments on the DiLiGenT dataset show that the automatically searched neural architectures performance compares favorably with the state-of-the-art uncalibrated PS methods.
  arXiv  Detail & Related papers  (2021-10-11T21:22:17Z)
• LCS: Learning Compressible Subspaces for Adaptive Network Compression at Inference Time [57.52251547365967]
  We propose a method for training a "compressible subspace" of neural networks that contains a fine-grained spectrum of models. We present results for achieving arbitrarily fine-grained accuracy-efficiency trade-offs at inference time for structured and unstructured sparsity. Our algorithm extends to quantization at variable bit widths, achieving accuracy on par with individually trained networks.
  arXiv  Detail & Related papers  (2021-10-08T17:03:34Z)
• Compact representations of convolutional neural networks via weight pruning and quantization [63.417651529192014]
  We propose a novel storage format for convolutional neural networks (CNNs) based on source coding and leveraging both weight pruning and quantization. We achieve a reduction of space occupancy up to 0.6% on fully connected layers and 5.44% on the whole network, while performing at least as competitive as the baseline.
  arXiv  Detail & Related papers  (2021-08-28T20:39:54Z)

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.