English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Citizen science is gaining popularity as a valuable tool for labelling largecollections of astronomical images by the general public. This is oftenachieved at the cost of poorer quality classifications made by amateurparticipants, which are usually verified by employing smaller data setslabelled by professional astronomers. Despite its success, citizen sciencealone will not be able to handle the classification of current and upcomingsurveys. To alleviate this issue, citizen science projects have been coupledwith machine learning techniques in pursuit of a more robust automatedclassification. However, existing approaches have neglected the fact that,apart from the data labelled by amateurs, (limited) expert knowledge of theproblem is also available along with vast amounts of unlabelled data that havenot yet been exploited within a unified learning framework. This paper presentsan innovative learning methodology for citizen science capable of takingadvantage of expert- and amateur-labelled data, featuring a transfer of labelsbetween experts and amateurs. The proposed approach first learns fromunlabelled data with a convolutional autoencoder and then exploits amateur andexpert labels via the pre-training and fine-tuning of a convolutional neuralnetwork, respectively. We focus on the classification of galaxy images from theGalaxy Zoo project, from which we test binary, multi-class, and imbalancedclassification scenarios. The results demonstrate that our solution is able toimprove classification performance compared to a set of baseline approaches,deploying a promising methodology for learning from different confidence levelsin data labelling.

CzSL: Learning from citizen science, experts and unlabelled data in astronomical image classification