mlsanityData to 7 Oct 2026

A Survey on Dynamic Neural Networks: from Computer Vision to Multi-modal Sensor Fusion

arXiv 2501.07451 ↗First posted 13 Jan 2025cs.CVcited by 5

Not on the map. Why?

Abstract (arXiv)

Model compression is essential in the deployment of large Computer Vision models on embedded devices. However, static optimization techniques (e.g. pruning, quantization, etc.) neglect the fact that different inputs have different complexities, thus requiring different amount of computations. Dynamic Neural Networks allow to condition the number of computations to the specific input. The current literature on the topic is very extensive and fragmented. We present a comprehensive survey that synthesizes and unifies existing Dynamic Neural Networks research in the context of Computer Vision. Additionally, we provide a logical taxonomy based on which component of the network is adaptive: the output, the computation graph or the input. Furthermore, we argue that Dynamic Neural Networks are particularly beneficial in the context of Sensor Fusion for better adaptivity, noise reduction and information prioritization. We present preliminary works in this direction. We complement this survey with a curated repository listing all the surveyed papers, each with a brief summary of the solution and the code base when available: https://github.com/DTU-PAS/awesome-dynn-for-cv .

Key facts. Cited by 5 papers in the map (data through 7 Oct 2026). Builds on 87 of its 143 references that are in the map. First posted on arXiv 13 Jan 2025.

Builds on

87 of 143 references.

References · 87

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,082
  3. 2014Going Deeper with Convolutionscited by 4,231
  4. 2020Knowledge Distillation: A Surveycited by 1,578
  5. 2022Token Merging: Your ViT But Fastercited by 995
  6. 2021DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsificationcited by 849
  7. 2021Scaling Vision with Sparse Mixture of Expertscited by 740
  8. 2017BranchyNet: Fast Inference via Early Exiting from Deep Neural Networkscited by 591
  9. 2019Be Your Own Teacher: Improve the Performance of Convolutional Neural Networks via Self Distillationcited by 491
  10. 2021AdaViT: Adaptive Tokens for Efficient Vision Transformercited by 360
  11. 2017SkipNet: Learning Dynamic Routing in Convolutional Networkscited by 328
  12. 2021AdaViT: Adaptive Vision Transformers for Efficient Image Recognitioncited by 228
38 more
  1. 2023Token Merging for Fast Stable Diffusioncited by 220
  2. 2023A Survey on Deep Neural Network Pruning-Taxonomy, Comparison, Analysis, and Recommendationscited by 200
  3. 2017BlockDrop: Dynamic Inference Paths in Residual Networkscited by 185
  4. 2021Chasing Sparsity in Vision Transformers: An End-to-End Explorationcited by 181
  5. 2021Adaptive Token Sampling For Efficient Vision Transformerscited by 181
  6. 2021Evo-ViT: Slow-Fast Token Evolution for Dynamic Vision Transformercited by 175
  7. 2019Edge AI: On-Demand Accelerating Deep Neural Network Inference via Edge Computingcited by 150
  8. 2017Convolutional Networks with Adaptive Inference Graphscited by 147
  9. 2020Resolution Adaptive Networks for Efficient Inferencecited by 146
  10. 2021Not All Images are Worth 16x16 Words: Dynamic Transformers for Efficient Image Recognitioncited by 143
  11. 2023Token Fusion: Bridging the Gap between Token Pruning and Token Mergingcited by 123
  12. 2016Spatially Adaptive Computation Time for Residual Networkscited by 118
  13. 2021Split Computing and Early Exiting for Deep Learning Applications: Survey and Research Challengescited by 114
  14. 2020SPINN: Synergistic Progressive Inference of Neural Networks over Device and Cloudcited by 109
  15. 2020Learning Dynamic Routing for Semantic Segmentationcited by 79
  16. 2023Joint Token Pruning and Squeezing Towards More Aggressive Compression of Vision Transformerscited by 79
  17. 2021Adaptive Focus for Efficient Video Recognitioncited by 76
  18. 2020Why should we add early exits to neural networks?cited by 70
  19. 2015Conditional Deep Learning for Energy-Efficient and Enhanced Pattern Recognitioncited by 67
  20. 2017Hard Mixtures of Experts for Large Scale Weakly Supervised Visioncited by 67
  21. 2019Improved Techniques for Training Adaptive Deep Networkscited by 66
  22. 2023Dynamic Token Pruning in Plain Vision Transformers for Semantic Segmentationcited by 61
  23. 2023SparseViT: Revisiting Activation Sparsity for Efficient High-Resolution Vision Transformercited by 60
  24. 2017Dynamic Deep Neural Networks: Optimizing Accuracy-Efficiency Trade-offs by Selective Executioncited by 59
  25. 2016Network of Experts for Large-Scale Image Categorizationcited by 54
  26. 2023Multi-modal Gated Mixture of Local-to-Global Experts for Dynamic Image Fusioncited by 54
  27. 2023Revisiting Token Pruning for Object Detection and Instance Segmentationcited by 50
  28. 2022Beyond Attentive Tokens: Incorporating Token Importance and Diversity for Efficient Vision Transformerscited by 47
  29. 2018You Look Twice: GaterNet for Dynamic Filter Selection in CNNscited by 46
  30. 2022Learning to Weight Samples for Dynamic Early-exiting Networkscited by 44
  31. 2021Zero Time Waste: Recycling Predictions in Early Exit Neural Networkscited by 39
  32. 2024Scene Adaptive Sparse Transformer for Event-based Object Detectioncited by 39
  33. 2021AdaFocus V2: End-to-End Training of Spatial Dynamic Networks for Video Recognitioncited by 37
  34. 2021Dynamic Network Quantization for Efficient Video Inferencecited by 36
  35. 2017Choosing Smartly: Adaptive Multimodal Fusion for Object Detection in Changing Environmentscited by 34
  36. 2023Dynamic Perceiver for Efficient Visual Recognitioncited by 33
  37. 2019Dual Dynamic Inference: Enabling More Efficient, Adaptive and Controllable Deep Inferencecited by 31
  38. 2020HAPI: Hardware-Aware Progressive Inferencecited by 30

50 most cited of 87 All 87 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 3; 2026: 2.

5 papers.

  1. 2025Knowing When to Quit: Probabilistic Early Exits for Speech Separationcited by 2
  2. 2025T-TAMER: Provably Taming Trade-offs in ML Servingcited by 0
  3. 2025ClustViT: Clustering-based Token Merging for Semantic Segmentationcited by 0
  4. 2026An Algorithm for On-Sensor Agnostic Detection of Changes in Human Activity for Ultra-Low-Power Applicationscited by 0
  5. 2026No Training, Better Flights: Test-Time Scaled VLMs for UAV Navigationcited by 0