Portugal & Friends 4 AI

AI4EU is a project funded by the EU’s H2020 program under GA-825619 and is developing a AI on-demand platform that allows communities, experts as well as AI-unskilled users, to create their own Artificial Intelligence systems in a simple and easy way. AI4EU combines a “social network” where users can communicate and share experiences and knowledge, a catalog of AI assets and a platform that connects those assets into complex AI pipelines that will be deployed with one click! To know more about its vision and strategy see The AI4EU Project sidebar.

This AI4EU NCP@Portugal‘s website documents community-building initiatives to disseminate the project and engage new participants beyond the project’s partners. The goal is to promote the shared effort of developing an open platform that can help overcoming the fragmented European AI landscape. These features are specially important for countries that lack critical mass of highly qualified workforce in ICT’s, such as Portugal.

We build on AI4EU – Work Package 7, in particular the Physical-AI task (T7.5) as well as the AI4IoT Pilot (Task 6.8 of WP6) . This team is enlarged with IST students and faculty who “volunteer” to develop new assets, allowing everybody to use them. This is a bottom-up initiative and if you wish to collaborate contact us filling the form .

Building AI Pipelines Automatically

AI4EU provides a catalog of assets that are connected and orchestrated in a ACUMOS design studio platform as illustrated below. Each of the “processing” components a “container”, with an environment totally isolated from every other container. In other words each asset is developed and runs in a sort of “virtual machine” and independent of every other “container”. To know more about ACUMOS and AI4EU Experiments see Martin Weiss’s tutorial on Youtube. Anyone can register and use the platform.

We’re not there yet !

If available, such tool is a “game changer” in many contexts. One obvious route is to enhance collaborative work between different scientific disciplines or teams that use different “tools” or environments, thus lowering the barrier of the “learning curve”! Another key area is technology transfer, where universities may be able to contribute with a special part or entire systems that can be shipped with our own tools. This allows collaboration with companies that lack specific skills or resources to pay the cost of implementing them. Finally, this tools can be the component that was missing for a more fruitful cooperation with developing countries.

However the final goal of “one click ” deployment, allowing experts and non-experts alike to assemble “customized” AI systems is not yet fully operational. Despite the progress with the ACUMOS platform, It is still a matter of research and development within AI4EU.

The same destination with a temporary detour

To start moving in this direction, in the context of AI4EU task T7.8, we developed a “manual” pipeline building and orchestrator system to simplify the assembly and deployment. We can start creating assets and pipelines right away and when the final platform reaches the production phase the assets can be seamlessly composed and deployed. The strategy is to add a bit more complexity to each component to gain simplicity in assembly/deployment.

This is to say that “you can start collaborating and leverage on community work right away”. See the Resources page to know how it works and what you can do with it.

How to navigate in this website

This is a “bottom-up” initiative, and we are at the very bottom of it! The website has a set of examples and categories of potential contributions, for which teechnical expertise and leadership are needed to push the idea forward:

• Infrastructure (distributed systems) development: Improving our orchestrator, interfaces and local ACUMOS install/development. Development of special modules (video streaming, interface with Android, ROS, visualization).
• Teaching : developing pipelines and tutorials for courses, labs and thesis support.
• Research: Production of special components (algorithms) that can be used by the research and academic community.
• Special Projects: Systems or subsystems on which “shoulders” others will stand to become giants. See examples on the sidebar

To know more or to discuss ideas and collaborations contact :