French research in Artificial Intelligence (AI) is recognized on the international scene for its scientific output.
Paradoxically, the number of software, visible and accessible, from French IT laboratories remains quite low.
An astonishing observation!
Research, like IT, advocates collaboration, sharing and pooling.
The IT tools and resources of a community of IT researchers is one of their major productions.
However, few codes are shared and pooled. AI researchers spend a lot of their time programming, often starting from scratch.
We often redevelop what has already been developed by others.
Researchers do not have time to take the step and convert their research prototypes. They often have to move on to new short-term funded projects, without being able to capitalize on what has been developed.
Thus, the transition from an experimental system that allows for experiments and publications, to a software tool accessible, documented and usable by others, is often left behind.
This is also explained by the funding of short-term projects, the lack of engineers and infrastructure in the laboratories.
French researchers and laboratories would benefit from making the most of the tools they develop. They would thus contribute to national and international visibility and influence.
Likewise, referencing and making available the tools developed by AI researchers could benefit different scientific communities, the world of higher education and companies of all sizes.
Inspired by the academic research project such as the project MultiTal carried out within the ERTIM laboratory of INALCO, PostLab is the collaborative laboratory of Artificial Intelligence (AI).
It aims to share, pool and enhance the tools developed in AI laboratories with the aim of simplifying access and use of this software.
Working in synergy with laboratory development organizations makes it possible, on the one hand, to speed up the work of AI researchers, increase their visibility and the influence of their laboratories; and on the other hand, to facilitate the use of AI tools for researchers, scientists and teachers in other disciplines.
Levels 4 to 7 represent the transition from experimental evidence in the laboratory to validation in a real environment. These steps, which are essential for transmitting innovation to manufacturers and to society, are often risky and expensive to take. This is why this path is called the “valley of death”.
The problem is very present in AI research. Indeed, very little software developed in the laboratory benefits industry and society.
This is where PostLab comes in to increase the TRL of software and bring it as far up the scale as possible. PostLab thus assists AI laboratories in making their software available and supports users in choosing the most relevant software for their needs.
The ambition of PostLab is to lower the barrier to entry of AI to benefit as many people as possible from the natural intelligence of artificial intelligence researchers.