Recent Posts

Archives

About us

Context and challenges facing research in Artificial Intelligence within academia

France’s research community production in AI is internationally acknowledged. Hence, one could expect the thriving community’s work to be equally as important as the number of software developed in the laboratories, yet those figures remain comparatively rather low. At first glance this state of affairs may seem surprising, as computer researchers favour cooperative models, knowledge sharing and resource pooling. All the more so when one would easily imagine the primary focus of a research computing team to be the production of software tools and resources.

Researchers spend a significant amount of time coding and programming, often starting from scratch as very few codes, just like the tools and resources used in research, are shared or pooled together. In many cases researchers (be they PhD students, post-docs or scientists) end up developing and redeveloping codes that have already been developed by fellow researchers. Indeed, one of the crucial issues here is the lack of time and resources to transform a code which is a prototype into that of a code of production quality. Short-term funding of research projects means they have to move on from one project to the next without being given the chance to promote or further develop what has been created and which most often gets left behind.

This situation can be explained to varying degrees by the issues at play in the current academic context: the lack of sustainable long term planning horizon, the want of engineers working in the laboratories, as well as the assessment of researchers on the basis of paper publication. The consequence of the general lack of funding leavs little room for the actual research activities to be conducted in the laboratory, as time and energy gets scattered across peripheral activities to attract grants and keep the actual structure up and running.

Yet, by taking advantage of the tools developed in French laboratories, universities and researchers developing AI software would gain visibility and impact at a national and international level. Likewise, the benefits of referencing and making the software easily accessible to everyone, could extend to various users including scientific communities with diverse interests, the higher education community and more broadly institutions,  individual users, associations and firms of diffrenet sizes.

We offer a collaborative project laboratory for the sharing, resource pooling and fostering of AI tools developed within academic laboratories

Its main purpose is to facilitate access and use of these tools. With this in mind, we’ve imagined an online platform based on the MultiTal research project developed by INALCO’s ERTIM laboratory. This platform draws on the latest technologies in the field of semantic web to easily reference those existing tools, while allowing for the generation of automatic multilingual documentation and the up-coming online use of these tools. Our membership platform is aimed at companies, laboratories and universities wishing to integrate software developed within Academia in their research and teaching practices. Training courses are also provided to end-users, ensuring that best use is made of these software tools. Consultancy services also provide guidance and analysis when it comes to choosing AI tools as well as building entrusted partnerships. Ultimately we strive to build a strong foundation for collaborations and partnerships between AI research teams, firms and laboratories from other disciplines.

Our intention is, in a way, to work together with the knowledge transfer, innovation and partner collaboration services in universities. This would enable, on the one hand, to forward the work being conducted by AI researchers, by putting their contribution in the spotlight and enhancing the impact of the different research teams. On the other hand, it would further assist in making these tools more easily obtainable amongst researchers, scientists and teachers across different fields. In order to reach these goals we offer:

1. referencing and mapping of IA tools

2. secure and reliable long-term hosting

3. an easy and intuitive access

4. a structured, precise and exhaustive multilingual and structured documentation

5. an integrated online software environment

6. a development environment that favours resource-pooling

7. supporting the use of these tools amongst different end-users, be they companies, academic members or students.