2026-07-13 早稲田大学
＜関連情報＞
- https://www.waseda.jp/inst/research/news/85155
- https://www.sciencedirect.com/science/article/pii/S0893608026006957
FracNeXt: フラクタルウェーブレットを用いたシーケンスモデルにおける視覚表現学習の強化 FracNeXt: Enhancing visual representation learning in sequence models with fractal wavelets
Pengfeng Lu, Sei-ichiro Kamata, Mengyunqiu Zhang
Neural Networks Available online: 7 June 2026
DOI:https://doi.org/10.1016/j.neunet.2026.109234
Highlights
- Fractal design enables robust and scalable model expansion.
- Wavelet patching enables spatially lossless token extraction.
- Wavelet connection ensures frequency-stable signal propagation.
Abstract
Recent applications of sequence models, such as Transformers and Mamba, in the vision domain have demonstrated promising performance. However, the processes of patch token extraction and token mixing inevitably result in the irreversible loss of spatial information. Existing sequence models for vision tasks rely heavily on residual connections to mitigate this issue, yet they overlook the limitations of residual connections in maintaining frequency stability. To address these challenges, we propose Multiple Wavelet Patch Partition (MWPP), a method that extracts patch tokens while preserving the spatial information within each patch. In addition, we introduce a frequency-aware Selective Wavelet Connection (SWC) to augment residual connections, thereby enhancing frequency stability and compensating for the information loss caused by token mixing. Building on MWPP and SWC, we design FracNeXt, a scalable fractal architecture that integrates both convolution and self-attention as token mixers. Under comparable experimental settings, FracNeXt achieves top-1 accuracies of 76.8% on ImageNet and 81.2% on CIFAR-100. Moreover, it delivers state-of-the-art performance across a variety of tasks, including object detection, optical character recognition, and time-series classification on diverse benchmarks. Furthermore, MWPP improves the F1 score of existing sequence models by up to 3.8%, while the proposed fractal architecture with SWC demonstrates superior robustness with respect to model depth.
