Group-Shared Low-Rank Approximation for Mobile-Efficient Pointwise Convolutions in Large-Kernel CNNs

• URL: http://arxiv.org/abs/2608.26069v2
• Date: Thu, 27 Aug 2026 08:17:51 GMT
• Title: Group-Shared Low-Rank Approximation for Mobile-Efficient Pointwise Convolutions in Large-Kernel CNNs
• Abstract summary: We propose Channel Group-Shared (CGS) low-rank approximation, a novel Singular Value Decomposition (SVD)-based parameter-sharing strategy.<n>CGS enables the feasible deployment of pre-trained CNN models on edge devices, thereby bridging the gap between high-performance vision models and practical edge deployment.
• Score: 44.54167094821349
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large-kernel Convolutional Neural Networks (CNNs) deliver remarkable performance in vision tasks by significantly expanding receptive fields, yet their quadratic parameter growth critically impedes storage-efficient edge deployment. While existing efficient architectures adopt parameter-efficient depthwise separable convolution backbones that leverage techniques like low-rank approximation and weight sharing to compress depthwise convolutions, we identify a critical oversight: pointwise convolutions dominate parameter volume (>87% in models like RepLKNet-31B) and constitute the primary deployment bottleneck on resource-constrained edge devices. This results in prohibitive storage costs and severe memory-loading constraints on resource-limited devices (e.g., smartphones with 4-12 GB Random Access Memory (RAM)). To overcome this, we propose Channel Group-Shared (CGS) low-rank approximation, a novel Singular Value Decomposition (SVD)-based parameter-sharing strategy. CGS constructs a structured low-rank paradigm isomorphic to SVD decomposition, comprising shared (high-parameter-cost) down/up-projection matrices across channel groups within a layer and channel-group-specific (low-parameter-cost) scalable diagonal matrices. This group-sharing design achieves significant parameter reduction. Extensive experiments demonstrate that large-kernel CNNs (RepLKNet, ConvNeXt, SLaK) enhanced with CGS strike an empirically favorable balance between competitive performance and substantially reduced storage costs. Crucially, by alleviating storage constraints, reducing memory bandwidth pressure during loading, and minimizing model loading latency, CGS enables the feasible deployment of pre-trained large-kernel CNN models on edge devices, thereby bridging the gap between high-performance vision models and practical edge deployment.

Related papers

• ACA-GS: Adaptive-Capacity Anchored Gaussian Splatting for Compact Dynamic Radiance Fields [29.09208632865348]
  We introduce an adaptive-capacity anchor-based framework that allocates representational capacity based on localtemporal demands.<n>Our method achieves up to 1.5x higher compression than state-of-the-art anchor-based methods while preserving comparable quality.
  arXiv  Detail & Related papers  (2026-08-05T08:47:48Z)
• Robust Auto-associative Memory via Convolutional Restricted Hopfield Networks [0.22940141855172036]
  Associative memory models play a fundamental role in pattern retrieval.<n>Existing approaches, including Modern Hopfield Networks (MHNs), and Predictive Coding Networks (PCNs) exhibit limitations in balancing storage capacity, computational efficiency, and robustness.<n>We propose a Convolutional Restricted Hopfield Networks (CRHNs), which integrates convolutional feature extraction with attractor-based memory retrieval in a structured latent space.
  arXiv  Detail & Related papers  (2026-06-11T21:25:54Z)
• LASER: Loss-Aware Singular-value Decomposition and Rank Allocation for Efficient Low-Precision Vision-Language Models [12.483003654711203]
  Vision-language models (VLMs) deliver strong multimodal reasoning capabilities, but their large computational cost and high parameter counts make deployment challenging on resource-constrained devices.<n>We proposetextitER (textbfLoss-textbfLASAware textbfSingular-value dtextbfEcomposition and textbfRank allocation), a low-rank compression framework for efficient low-precision VLM inference.
  arXiv  Detail & Related papers  (2026-05-30T06:53:23Z)
• OScaR: The Occam's Razor for Extreme KV Cache Quantization in LLMs and Beyond [50.440302567029654]
  Multi-modal intelligence has made the Key-Value cache a dominant memory bottleneck for efficient deployment.<n>In this work, we revisit the inherent limitations of the per-channel quantization paradigm.<n>We propose OScaR, an accurate and lightweight KV cache compression framework for X-LLMs.
  arXiv  Detail & Related papers  (2026-05-19T10:53:03Z)
• Ge$^\ ext{2}$mS-T: Multi-Dimensional Grouping for Ultra-High Energy Efficiency in Spiking Transformer [84.8831358775386]
  Spiking Neural Networks (SNNs) offer superior energy efficiency over Artificial Neural Networks (ANNs)<n>Existing paradigms including ANN-SNN Conversion and Spatial-Temporal Backpropagation (STBP) suffer from inherent limitations.<n>We propose Ge$text2$mS-T, a novel architecture implementing grouped computation across temporal, spatial and network structure dimensions.
  arXiv  Detail & Related papers  (2026-04-10T02:58:46Z)
• HPC: Hierarchical Point-based Latent Representation for Streaming Dynamic Gaussian Splatting Compression [14.759497852655047]
  We propose HPC, a novel streaming dynamic Gaussian Splatting compression framework.<n>It employs a hierarchical point-based latent representation that operates on a per-Gaussian basis to avoid parameter redundancy in unoccupied space.<n>It achieves a storage reduction of 67% against its baseline while maintaining high reconstruction fidelity.
  arXiv  Detail & Related papers  (2026-01-31T11:35:02Z)
• SHERL: Synthesizing High Accuracy and Efficient Memory for Resource-Limited Transfer Learning [63.93193829913252]
  We propose an innovative METL strategy called SHERL for resource-limited scenarios. In the early route, intermediate outputs are consolidated via an anti-redundancy operation. In the late route, utilizing minimal late pre-trained layers could alleviate the peak demand on memory overhead.
  arXiv  Detail & Related papers  (2024-07-10T10:22:35Z)
• Learning k-Level Structured Sparse Neural Networks Using Group Envelope Regularization [4.0554893636822]
  We introduce a novel approach to deploy large-scale Deep Neural Networks on constrained resources. The method speeds up inference time and aims to reduce memory demand and power consumption.
  arXiv  Detail & Related papers  (2022-12-25T15:40:05Z)
• Efficient Micro-Structured Weight Unification and Pruning for Neural Network Compression [56.83861738731913]
  Deep Neural Network (DNN) models are essential for practical applications, especially for resource limited devices. Previous unstructured or structured weight pruning methods can hardly truly accelerate inference. We propose a generalized weight unification framework at a hardware compatible micro-structured level to achieve high amount of compression and acceleration.
  arXiv  Detail & Related papers  (2021-06-15T17:22:59Z)
• Adaptive Subcarrier, Parameter, and Power Allocation for Partitioned Edge Learning Over Broadband Channels [69.18343801164741]
  partitioned edge learning (PARTEL) implements parameter-server training, a well known distributed learning method, in wireless network. We consider the case of deep neural network (DNN) models which can be trained using PARTEL by introducing some auxiliary variables.
  arXiv  Detail & Related papers  (2020-10-08T15:27:50Z)

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.