The Fellowship of Online Machine Learning (FOML) is a loose community of practitioners, researchers, and people in the industry. Our goal is to promote online machine learning through software, research publications, learning material, discussions, and events. We are also here to advise companies looking to do online machine learning in practice.
Online machine learning encompasses algorithms which consume streaming data. An online ML algorithm infers and learns one sample at a time.
It is different to traditional machine learning, which is usually performed in a batch manner. In fact, online machine learning is a completely different paradigm.
By reducing the gap between training and inference, online machine learning can be easier to put into production, and may be more natural to work with. We are convinced online machine learning is more adequate than batch machine learning in certain situations.
We believe online machine learning hasn’t reached the popularity it deserves. We think this is due to several reasons:
Our (ambitious) goal is to address these three gaps.
We have a Discord server revolving around online machine learning. Feel free to join, introduce yourself, and ask/suggest anything that goes through your mind!
We encourage you to use GitHub Discussions and GitHub Issues for code-related discussions. That way, if you ask a question, other people will be able to find the answer too.
If none of these communication channels feel adequate, feel free to reach out to [email protected].
A Python library for building online machine learning models.
A side-car to https://github.com/online-ml/river with a lower barrier of acceptance for contributions.