mlsanityData to 14 Oct 2026

Compression of Deep Convolutional Neural Networks for Fast and Low Power Mobile Applications

arXiv 1511.06530 ↗First posted 20 Nov 2015cs.CV cs.LGcited by 162

Not on the map. Why?

Abstract (arXiv)

Although the latest high-end smartphone has powerful CPU and GPU, running deeper convolutional neural networks (CNNs) for complex tasks such as ImageNet classification on mobile devices is challenging. To deploy deep CNNs on mobile devices, we present a simple and effective scheme to compress the entire CNN, which we call one-shot whole network compression. The proposed scheme consists of three steps: (1) rank selection with variational Bayesian matrix factorization, (2) Tucker decomposition on kernel tensor, and (3) fine-tuning to recover accumulated loss of accuracy, and each step can be easily implemented using publicly available tools. We demonstrate the effectiveness of the proposed scheme by testing the performance of various compressed CNNs (AlexNet, VGGS, GoogLeNet, and VGG-16) on the smartphone. Significant reductions in model size, runtime, and energy consumption are obtained, at the cost of small loss in accuracy. In addition, we address the important implementation level issue on 1?1 convolution, which is a key operation of inception module of GoogLeNet as well as CNNs compressed by our proposed scheme.

Key facts. Cited by 162 papers in the map (data through 14 Oct 2026). First posted on arXiv 20 Nov 2015.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 46; 2022: 32; 2023: 31; 2024: 23; 2025: 16; 2026: 14.

162 papers.

Sort:

By area:

  1. 2023ASVD: Activation-aware Singular Value Decomposition for Compressing Large Language Modelscited by 169
  2. 2024Task Singular Vectors: Reducing Task Interference in Model Mergingcited by 147
  3. 2023Structured Pruning for Deep Convolutional Neural Networks: A surveycited by 122
  4. 2025No Task Left Behind: Isotropic Model Merging with Common and Task-Specific Subspacescited by 96
  5. 2023Multi-Grid Tensorized Fourier Neural Operator for High-Resolution PDEscited by 68
  6. 2022A Comprehensive Survey on Model Quantization for Deep Neural Networks in Image Classificationcited by 65
  7. 2021Tensor Methods in Computer Vision and Deep Learningcited by 52
  8. 2021Towards Efficient Tensor Decomposition-Based DNN Model Compression with Optimization Frameworkcited by 52
  9. 2021A Survey on Green Deep Learningcited by 52
  10. 2022Low-rank lottery tickets: finding efficient low-rank neural networks via matrix differential equationscited by 43
  11. 2024Tiny Machine Learning: Progress and Futurescited by 40
  12. 2021Towards Compact CNNs via Collaborative Compressioncited by 36
38 more
  1. 2021Revisiting Dynamic Convolution via Matrix Decompositioncited by 35
  2. 2021Compressing Neural Networks: Towards Determining the Optimal Layer-wise Decompositioncited by 33
  3. 2022Enable Deep Learning on Mobile Devices: Methods, Systems, and Applicationscited by 32
  4. 2023Tensor Networks Meet Neural Networks: A Survey and Future Perspectivescited by 32
  5. 2021ZerO Initialization: Initializing Neural Networks with only Zeros and Onescited by 28
  6. 2022Data-Efficient Structured Pruning via Submodular Optimizationcited by 23
  7. 2021LegoDNN: Block-grained Scaling of Deep Neural Networks for Mobile Visioncited by 20
  8. 2022Pruning Neural Networks via Coresets and Convex Geometry: Towards No Assumptionscited by 20
  9. 2023HALOC: Hardware-Aware Automatic Low-Rank Compression for Compact Neural Networkscited by 20
  10. 2021NanoNet: Real-Time Polyp Segmentation in Video Capsule Endoscopy and Colonoscopycited by 16
  11. 2022Compression-aware Training of Neural Networks using Frank-Wolfecited by 16
  12. 2024CoMERA: Computing- and Memory-Efficient Training via Rank-Adaptive Tensor Optimizationcited by 15
  13. 2021Layer Folding: Neural Network Depth Reduction using Activation Linearizationcited by 14
  14. 2021Convolutional Neural Network Compression through Generalized Kronecker Product Decompositioncited by 14
  15. 2023Tensor Decomposition for Model Reduction in Neural Networks: A Reviewcited by 14
  16. 2023A Survey on Approximate Edge AI for Energy Efficient Autonomous Driving Servicescited by 14
  17. 2023COMCAT: Towards Efficient Compression and Customization of Attention-Based Vision Modelscited by 14
  18. 2022Towards Lightweight Black-Box Attacks against Deep Neural Networkscited by 13
  19. 2023Geometry-aware training of factorized layers in tensor Tucker formatcited by 12
  20. 2024LoRETTA: Low-Rank Economic Tensor-Train Adaptation for Ultra-Low-Parameter Fine-Tuning of Large Language Modelscited by 12
  21. 2022A Passive Similarity based CNN Filter Pruning for Efficient Acoustic Scene Classificationcited by 11
  22. 2024ELRT: Efficient Low-Rank Training for Compact Convolutional Neural Networkscited by 11
  23. 2021Joint Matrix Decomposition for Deep Convolutional Neural Networks Compressioncited by 10
  24. 2023GOHSP: A Unified Framework of Graph and Optimization-based Heterogeneous Structured Pruning for Vision Transformercited by 10
  25. 2021Manipulating Identical Filter Redundancy for Efficient Pruning on Deep and Complicated CNNcited by 9
  26. 2021Resource-Efficient Deep Learning: A Survey on Model-, Arithmetic-, and Implementation-Level Techniquescited by 9
  27. 2022Towards Practical Control of Singular Values of Convolutional Layerscited by 9
  28. 2021Towards Extremely Compact RNNs for Video Recognition with Fully Decomposed Hierarchical Tucker Structurecited by 8
  29. 2021Knowledge Distillation via Instance-level Sequence Learningcited by 8
  30. 2022Deformable Butterfly: A Highly Structured and Sparse Linear Transformcited by 8
  31. 2022A Unified Weight Initialization Paradigm for Tensorial Convolutional Neural Networkscited by 8
  32. 2022Inflating 2D Convolution Weights for Efficient Generation of 3D Medical Imagescited by 8
  33. 2023Approximately Optimal Core Shapes for Tensor Decompositionscited by 8
  34. 2023Adaptive Compression-Aware Split Learning and Inference for Enhanced Network Efficiencycited by 8
  35. 2025DipSVD: Dual-importance Protected SVD for Efficient LLM Compressioncited by 8
  36. 2022Wider or Deeper Neural Network Architecture for Acoustic Scene Classification with Mismatched Recording Devicescited by 7
  37. 2025Forget the Data and Fine-Tuning! Just Fold the Network to Compresscited by 7
  38. 2021Toward Compact Deep Neural Networks via Energy-Aware Pruningcited by 6

50 most cited of 162 All 162 (JSON)