Join us on the AI4EU Experiments platform: Find out how to build AI solutions in an intuitive way! For this, a visual editor is available in the design studio of the AI4EU Experiments platform, which is open to anyone. The visual design method facilitates the creation of human-centered AI-solutions, building modular structures and using hybrid AI technologies. We invite you to discover our resources, create your own building blocks, try out the design capabilities and let existing applications inspire and guide you on your way to developing your own solutions. You are welcome to collaborate and share your ideas and solutions with the AI community.
The visual editor of the AI4EU Experiments design studio allows building AI solutions, also called pipelines, in a quick and intuitive way. The studio provides the necessary tools for AI developers: a collection of building blocks for your solutions, a canvas to build on and visual guidance on how to connect the blocks. The AI4EU Experiments design studio is a part of the open source AI4EU Experiments platform, running on a customized Acumos instance, a platform and open source framework. The platform supports the development of AI solutions as portable, containerized micro-services. Users are free to choose any modelling language, toolkit, run-time infrastructure or cloud service. The studio is open to anyone, you just have to create a user account and sign in.
Create AI modules
The AI4EU Experiments platform facilitates the creation of modular AI solutions. When building your pipeline, you can choose from the re-usable, existing building blocks, for example AI models for object recognition, classification or segmentation. To manage the data flow, select from available data brokers and add splitters or collators if necessary. We invite users to create and upload their own building blocks to the AI resources catalogue and use these in the design studio as well. Anyone who has registered and created a user account for the AI4EU Experiments platform can upload resources, which are checked for quality by our experts and then published in the resources catalogue.
Once your solution is ready, we recommend onboarding it as a containerized micro-service, for example as docker container. It is not even necessary to upload the container; instead, you can provide a link referencing the storage facility, such as a docker registry. In that way, users can easily access the solution and deploy it on any execution environment.
Build hybrid AI-solutions
The modular design concept helps with building hybrid AI-solutions. Such hybrid solutions combine different AI technologies, for example classic Machine Learning with symbolic AI, reasoning or constraint programming. Combining two or more AI methods is beneficial in some application scenarios, especially when data sets are incomplete or the application demands highest standards for reliability, explainability or trustworthiness.
A key prerequisite for re-using and flexibly combining AI building blocks based on different AI technologies is the specification of standardized input and output data interfaces. Future research and development projects on the AI4EU Experiments platform will deal with these questions, thus prototyping the interaction of different AI technologies.
Learn from examples
Let’s now look at an example for an AI pipeline built with the AI4EU Experiments design studio. The screen shots below illustrate the process of designing a video pipeline. The chosen building blocks are on the design canvas, other available resources listed on the left.
The main building blocks of the solution are two models: the video segmentation and the video object recognition. A video file broker manages the data input. This basic pipeline is further enhanced by auxiliary infrastructure nodes depicted on a purple background: a persistent volume provider on the left and a model initializer. Finally, the developer added a tensor board connector, a tool for analyzing a model, and connected it to the model for video object recognition.
The design studio assists pipeline builders in identifying the right interfaces or ports for connecting the building blocks: when clicking on one source port, possible connecting ports flash. If the connection between two ports is technically correct, the developer gets the confirming message “It’s a match!”
The visual editor available on the AI4EU experiments platform thus helps to collaboratively create AI solutions and share them with the AI community. Re-usable building blocks facilitate modular designs and hybrid solutions. This is a step towards the development of human-centred AI, which meets the highest standards with respect to reliability, explainability and trustworthiness. We invite you to find out more, discover our resources, create your own building blocks and let existing solutions inspire and guide you to design your own. We welcome you to join this pioneering team!
Further information and resources:
- The AI4EU Experiments Platform: https://acumos-int-fhg.ai4eu.eu/#/home
- The AI4EU YouTube channel https://www.youtube.com/channel/UC0TvS1wLb1Qja9k-dle0cdQ has two videos with tutorials on pipeline building and one AI4EU Café session on the subject:
- Tutorial on building a basic pipeline with AI4EU Experiments: https://www.youtube.com/watch?v=vZXoEUTO2bQ
- Tutorial on building a video pipeline with additional infrastructure nodes: https://www.youtube.com/watch?v=AYMaVtiPuA4
- AI4EU Café session with Martin Welß “Composing AI Pipelines with AI4EU Experiments”: https://www.youtube.com/watch?v=hw1s225rhwQ
- Links to GitHub resources created by the AI4EU Experiments team:
- Specification documents for containers https://github.com/ai4eu/tutorials/tree/master/Container_Specification
- Tutorials https://github.com/ai4eu/tutorials
- Example for an AI audio pipeline created in the AI4EU Experiments design studio https://acumos-int-fhg.ai4eu.eu/#/marketSolutions?solutionId=df1f1286-0071-4df8-afd7-fe5dd20f9cd4&revisionId=69c2bcae-398d-46a0-9288-efe28137e6ef#md-model-detail-template
- About the Acumos AI platform and open source framework: https://www.acumos.org/
- AI4EU Experiments
- hybrid AI
- modular AI