# Resolution Allocation with Learnable Resizing and Spectral Filter Parameterization

arXiv: https://arxiv.org/abs/2610.20274 · HTML: https://mlsanity.com/paper/2610.20274/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 10 of its 17 references that are in the map.
- First posted on arXiv 14 Oct 2026.

## Builds on (most cited first, 10 of 10)

- [Implicit Neural Representations with Periodic Activation Functions](https://mlsanity.com/paper/2006.09661/) (2020)
- [DARTS: Differentiable Architecture Search](https://mlsanity.com/paper/1806.09055/) (2018)
- [CKConv: Continuous Kernel Convolution For Sequential Data](https://mlsanity.com/paper/2102.02611/) (2021)
- [Spectral Representations for Convolutional Neural Networks](https://mlsanity.com/paper/1506.03767/) (2015)
- [Learning strides in convolutional neural networks](https://mlsanity.com/paper/2202.01653/) (2022)
- [How Convolutional Neural Networks Deal with Aliasing](https://mlsanity.com/paper/2102.07757/) (2021)
- [Shape Adaptor: A Learnable Resizing Module](https://mlsanity.com/paper/2008.00892/) (2020)
- [Pooling Revisited: Your Receptive Field is Suboptimal](https://mlsanity.com/paper/2205.15254/) (2022)
- [DNArch: Learning Convolutional Neural Architectures by Backpropagation](https://mlsanity.com/paper/2302.05400/) (2023)
- [Scaling Continuous Kernels with Sparse Fourier Domain Learning](https://mlsanity.com/paper/2409.09875/) (2024)

## Abstract (arXiv)

Learnable downsampling and spectral-domain filter parameterizations have each made convolutional networks more flexible, but their interaction has not been systematically examined. We bridge this gap by examining convolutional kernels learned in the spatial and spectral domains, together with learned downsampling ratios under a cropped resizing implementation. We find that spectral and spatial kernels respond differently to learned resolution allocation, with larger gains for the spectral kernel. We observe this consistently on three datasets (EuroSAT, CIFAR-10 and STL-10) and at two model sizes (differences in accuracy gain of +0.013 to +0.028 on held-out test sets). Part of this difference reflects the greater gain of the spectral kernel from retaining more spatial area, but the difference remains positive when total retained area is matched. This pattern is observed across datasets and model sizes, although its magnitude and dependence on the complexity budget vary with model size and the resizing implementation.