# Recursive Autoconvolution for Unsupervised Learning of Convolutional Neural Networks

Source: [Recursive Autoconvolution for Unsupervised Learning of Convolutional Neural Networks](<https://www.sota2.com/research/paper/68fa63391e209131ce9fd907>)

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

Data updated: 2025-10-25T20:36:15.201Z

Authors: Boris Knyazev, Erhardt Barth, Thomas Martinetz

Published: 2016-06-02T10:37:46.000Z

## Abstract

In visual recognition tasks, such as image classification, unsupervised learning exploits cheap unlabeled data and can help to solve these tasks more efficiently. We show that the recursive autoconvolution operator, adopted from physics, boosts existing unsupervised methods by learning more discriminative filters. We take well established convolutional neural networks and train their filters layer-wise. In addition, based on previous works we design a network which extracts more than 600k features per sample, but with the total number of trainable parameters greatly reduced by introducing shared filters in higher layers. We evaluate our networks on the MNIST, CIFAR-10, CIFAR-100 and STL-10 image classification benchmarks and report several state of the art results among other unsupervised methods.

## 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/1606.00611>)

[PDF](<https://arxiv.org/pdf/1606.00611.pdf>)