# Learnable Polyphase Sampling for Shift Invariant and Equivariant Convolutional Networks

Source: [Learnable Polyphase Sampling for Shift Invariant and Equivariant Convolutional Networks](<https://www.sota2.com/research/paper/68fa6dab1e209131cea3b532>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-01-04T05:22:45.709Z

Authors: Renan A. Rojas-Gomez, Teck-Yian Lim, Alexander G. Schwing, Minh N. Do, Raymond A. Yeh

Published: 2022-10-14T17:59:55.000Z

## Abstract

We propose learnable polyphase sampling (LPS), a pair of learnable down/upsampling layers that enable truly shift-invariant and equivariant convolutional networks. LPS can be trained end-to-end from data and generalizes existing handcrafted downsampling layers. It is widely applicable as it can be integrated into any convolutional network by replacing down/upsampling layers. We evaluate LPS on image classification and semantic segmentation. Experiments show that LPS is on-par with or outperforms existing methods in both performance and shift consistency. For the first time, we achieve true shift-equivariance on semantic segmentation (PASCAL VOC), i.e., 100% shift consistency, outperforming baselines by an absolute 3.3%.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2210.08001>)

[PDF](<https://arxiv.org/pdf/2210.08001.pdf>)

[Implementation](<https://raymondyeh07.github.io/learnable_polyphase_sampling/>)