Monday webinar - getting machine learning models into use

To make machine learning models for analysis of biological data valuable, a number of links in the value chain have to be in place: data gathering, data cleaning and pre-processing, training of the model, and deployment of the model such that it is available for those who actually need to use it in their day-to-day work.

While we may tend think about machine learning projects as mainly being about gathering data and training the models, the down stream work required to actually deploy the trained models may be equally critical to a project's real life impact. This talk focuses how to deploy bioinformatics projects, to make sure that they get all the way to the finish line, and into the hands of the end-users that need them.

In the Webinar, Jeppe Hallgren, author of the Open Protein machine learning framework, will share his experiences with building and sharing machine learning algorithms and give a practical introduction to the BioLib deployment platform.

Various themes and challenges will be discussed including: scalability and compatibility across platforms; how to make tools useful for different scientific audiences with varying programming skills; data-security and protection of sensitive biological data; and the challenges around maintainability.