Simple-ML is a project funded by the German Federal Ministry of Education and Research (BMBF). Our goal is to increase the usability of creating machine learning workflows.
The efficient application of current Machine Learning (ML) procedures requires a very high level of expert knowledge, which stands in the way of a widespread use of ML approaches, especially by small and medium-sized enterprises. The goal of the Simple-ML project is to significantly improve the usability of ML processes in order to make them more accessible to a broad user group.
The definition of a domain-specific language (DSL) is a central contribution of the project: The DSL provides an integrated description of the ML workflows and their components, and can be specified by textual and graphical editors. Furthermore, the project contributes towards the robustness of created ML workflows, explainability and transparency of the learned models, efficiency and scalability of the created applications, as well as towards the reusability of the created solutions. This is done by applying semantic technologies, extending symbolic ML procedures and by setting up scalable ML frameworks.
- Install Docker.
- On Linux: Install Docker Compose v2.
- Clone this repository.
- Install and start Simple-ML by running the following command in the root directory of the cloned repository. This will take a few minutes the first time you do it:
docker compose up
- Open localhost:4200 in a Chromium-based browser like Chrome or the new Edge. Other browser like Firefox or Internet Explorer are not yet supported.
- Later: After pulling again from the repository you need to update the Docker images:
docker compose build
The following page provides a tutorial for users of Simple-ML:
Developers get more information about the implementation of the specific components here: