Following changes are planned:

• Introduce MNM API and make publicly accessible via PyPi
• Introduce example workflows using open source datasets
• Introduce proper wiki and documentation pages
• Remove necessary in-code adjustment requirement for some types of experiments.
• Make platform data/covariate agnostic
• Add multidimensional data support (e.g. voxel based / tensor based data without underlying assumptions)
• Include easier workflows for univariate models
• Validate and/or delegate univariate models to the PCN toolkit
• Add support for multi GPU / HPC computing setups
• Define different analysis methods as drop-in modules which can be downloaded and shared independently from the platform
• Make default config easier to use
• Improve support for checkpointing and fine-tuning of exisiting models
• Integrate external data preprocessing and data harmonizations methods
• Improve out-of-the-box usage experience by making docker image publicly accessible with folder mounts to access configuration options.