HiRo: A Compact Four-Directional Hierarchical Reservoir Token-Mixer for Efficient Image Classification

• URL: http://arxiv.org/abs/2606.15151v1
• Date: Sat, 13 Jun 2026 06:44:06 GMT
• Title: HiRo: A Compact Four-Directional Hierarchical Reservoir Token-Mixer for Efficient Image Classification
• Abstract summary: HiRo is a parameter-efficient image classification model that integrates shifted-window partitioning and hierarchical reservoir computing.<n>Images are divided into non-overlapping patches (treated as tokens), linearly projected, normalized, and enriched with 2D sinusoidal positional encodings, then processed within local windows.<n>Despite using under 1M trainable parameters, HiRo also achieves 99.46%, 85.57%, and 59.10% accuracy on MNIST, CIFAR-10, and CIFAR-100, respectively.
• Score: 0.19116784879310025
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recent image classification models must balance local feature modeling, cross-window interaction, and parameter efficiency. Many high-performing architectures rely on fully trainable token-mixers, which improve representation learning but increase parameter count, optimization complexity and computational cost. We propose a parameter-efficient image classification model called HiRo that integrates shifted-window partitioning with multi-directional hierarchical reservoir computing. Images are divided into non-overlapping patches (treated as tokens), linearly projected, normalized, and enriched with 2D sinusoidal positional encodings, then processed within local windows. Inside each window, tokens are scanned in four directions and passed through a two-stage slice-and-mix reservoir module. In the first stage, directional sequences are split into contiguous slices, each processed by its own fixed reservoir with a trainable closed-loop readout. The resulting slice outputs are summarized using the start, end, and mean representations, and then mixed by a second-stage fixed reservoir for each direction. The mixed slice representations are expanded back to the token level and fused with the first-stage outputs, after which the four directional outputs are realigned and averaged. Consecutive blocks alternate between regular and shifted windows to enable cross-window interaction, followed by layer normalization, a residual feed-forward network, and global pooling for classification. This design combines regular and shifted window partitioning with hierarchical multi-directional reservoirs to make an efficient local-to-cross-window token-mixing framework for image classification. Despite using under 1M trainable parameters and significantly lower memory and time than transformer-style baselines, HiRo also achieves 99.46%, 85.57%, and 59.10% accuracy on MNIST, CIFAR-10, and CIFAR-100, respectively.

Related papers

• Latent Recurrent Transformer: Architecture Exploration, Training Strategies, and Scaling Behavior [107.2098567818173]
  Latent Recurrent Transformer (LRT) is a lightweight augmentation of autoregressive transformers.<n>LRT reuses a high-level source-layer hidden state from the previous token as recurrent memory for the next token.
  arXiv  Detail & Related papers  (2026-05-26T10:10:26Z)
• Covariance-Aware Goodness for Scalable Forward-Forward Learning [8.12004710498018]
  In convolutional settings, Forward-Forward methods significantly under-perform backpropagation on complex benchmarks.<n>We propose a framework centered on three key components.<n>The resulting BP-free model achieves 73.01% on ImageNet-100 and 50.30% on Tiny-ImageNet.
  arXiv  Detail & Related papers  (2026-05-05T23:08:00Z)
• GSPN-2: Efficient Parallel Sequence Modeling [101.33780567131716]
  Generalized Spatial Propagation Network (GSPN) addresses this by replacing quadratic self-attention with a line-scan propagation scheme.<n>GSPN-2 establishes a new efficiency frontier for modeling global spatial context in vision applications.
  arXiv  Detail & Related papers  (2025-11-28T07:26:45Z)
• Programmable k-local Ising Machines and all-optical Kolmogorov-Arnold Networks on Photonic Platforms [0.0]
  Photonic computing promises energy-efficient acceleration for optimization and learning.<n>Here we unify k-local Ising optimization and optical Kolmogorov-Arnold network (KAN) learning on a single photonic platform.<n>Key idea is to convert the structural nonlinearity of a nominally linear scatterer into a per-window computational resource.
  arXiv  Detail & Related papers  (2025-08-24T16:39:09Z)
• FractMorph: A Fractional Fourier-Based Multi-Domain Transformer for Deformable Image Registration [0.6683923149620578]
  We present FractMorph, a novel 3D dual-parallel transformer-based architecture that enhances cross-image feature matching.<n>A lightweight U-Net style network then predicts a dense deformation field from the transformer-enriched features.<n>Results show FractMorph achieves state-of-the-art performance with an overall Dice Similarity Coefficient (DSC) of $86.45%$, an average per-structure of $75.15%$, and a 95th-percentile Hausdorff distance (HD95) of $1.54mathrmmm$ on our data split.
  arXiv  Detail & Related papers  (2025-08-17T17:42:10Z)
• SepFormer: Coarse-to-fine Separator Regression Network for Table Structure Recognition [0.5120567378386615]
  We present SepFormer, which integrates the split-and-merge paradigm into a single step through separator regression with a DETR-style architecture.<n>SepFormer can run on average at 25.6 FPS while achieving comparable performance with state-of-the-art methods on several benchmark datasets.
  arXiv  Detail & Related papers  (2025-06-27T05:20:42Z)
• Attamba: Attending To Multi-Token States [6.5676809841642125]
  We introduce Attamba, a novel architecture that uses state-space models to compress chunks of tokens. We find that replacing key and value projections in a transformer with SSMs can improve model quality and enable flexible token chunking. Attamba can perform attention on chunked-sequences of variable length, enabling a smooth transition between quadratic and linear scaling.
  arXiv  Detail & Related papers  (2024-11-26T18:52:06Z)
• Multiway Point Cloud Mosaicking with Diffusion and Global Optimization [74.3802812773891]
  We introduce a novel framework for multiway point cloud mosaicking (named Wednesday) At the core of our approach is ODIN, a learned pairwise registration algorithm that identifies overlaps and refines attention scores. Tested on four diverse, large-scale datasets, our method state-of-the-art pairwise and rotation registration results by a large margin on all benchmarks.
  arXiv  Detail & Related papers  (2024-03-30T17:29:13Z)
• ParFormer: A Vision Transformer with Parallel Mixer and Sparse Channel Attention Patch Embedding [9.144813021145039]
  This paper introduces ParFormer, a vision transformer that incorporates a Parallel Mixer and a Sparse Channel Attention Patch Embedding (SCAPE) ParFormer improves feature extraction by combining convolutional and attention mechanisms. For edge device deployment, ParFormer-T excels with a throughput of 278.1 images/sec, which is 1.38 $times$ higher than EdgeNeXt-S. The larger variant, ParFormer-L, reaches 83.5% Top-1 accuracy, offering a balanced trade-off between accuracy and efficiency.
  arXiv  Detail & Related papers  (2024-03-22T07:32:21Z)
• Dynamic Frame Interpolation in Wavelet Domain [57.25341639095404]
  Video frame is an important low-level computation vision task, which can increase frame rate for more fluent visual experience. Existing methods have achieved great success by employing advanced motion models and synthesis networks. WaveletVFI can reduce computation up to 40% while maintaining similar accuracy, making it perform more efficiently against other state-of-the-arts.
  arXiv  Detail & Related papers  (2023-09-07T06:41:15Z)
• Green Hierarchical Vision Transformer for Masked Image Modeling [54.14989750044489]
  We present an efficient approach for Masked Image Modeling with hierarchical Vision Transformers (ViTs) We design a Group Window Attention scheme following the Divide-and-Conquer strategy. We further improve the grouping strategy via the Dynamic Programming algorithm to minimize the overall cost of the attention on the grouped patches.
  arXiv  Detail & Related papers  (2022-05-26T17:34:42Z)
• Augmenting Convolutional networks with attention-based aggregation [55.97184767391253]
  We show how to augment any convolutional network with an attention-based global map to achieve non-local reasoning. We plug this learned aggregation layer with a simplistic patch-based convolutional network parametrized by 2 parameters (width and depth) It yields surprisingly competitive trade-offs between accuracy and complexity, in particular in terms of memory consumption.
  arXiv  Detail & Related papers  (2021-12-27T14:05:41Z)

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.