Infrastructure for creation of shared image & sound recognition algorithms

This document provides an overview of the online available infrastructure for training and deploying image and sound recognition models with a focus on species identification. The main document is split in two parts, one part focussing on the construction of sound recognition models and one part focussed on the construction of image recognition models. The field of machine learning is in full development and new tools are becoming available regularly while others become outdated and obsolete. For this reason, the text of this report will be published as two living documents so that it can serve as a starting point for a new generation of AI-biodiversity researchers.