The overall scope of the journal is to publish original papers on:

• studies on the design and behaviour of learning in intelligent systems;
• definition of new learning tasks and methods of assessment;
• development of actual algorithms in presence of data validation;
• evidence of successful application in any of the four above fundamental sciences.

Further, it also hosts a repository of ML open source software collecting implementations of non-trivial ML algorithms, numerical toolboxes, etc. for scientific computing.

Finally, JMLFS also publishes proceedings of ML conferences, workshops, etc. as special issues.