Following on from the previous article, which focused on the significant potential for companies to benefit from the proposals of the Villani report on artificial intelligence (AI), the present article aims to examine one of these proposals in greater depth: "for agile and disseminating research", and more specifically "to bring together researchers, students and companies".
This theme can be likened to a buzzword that has been heard for several decades. It corresponds to the constant concern of certain political decision-makers to build bridges between research that is too theoretical and companies that lack innovation due to the presumed inability of the majority of them to develop fundamental research. This observation is partially contradicted by the substantial investments made by GAFA* in fundamental research centers with very significant resources, but which have a perverse effect: the gradual drying up of public research with many of the best students deciding to leave permanently for the private sector. The report's proposal is therefore based on a public/private collaboration.
The Villani report insists on the idea of innovation based on this cooperation. Companies could thus regularly discuss their business problems with researchers, and thus encourage the evolution of their solutions according to the lessons they receive from researchers. These exchanges could, for example, focus on the optimization of their algorithms or on audits of theirAI systems.
What is the 3IA?
To support this axiom, the Villani report recommends the creation of what it calls 3IAs, i.e. a network of interdisciplinary artificial intelligence institutes. These institutes, numbering 5-6, would be distributed throughout the country and would compensate for the dispersion noted by the report of the various existing research institutes (CNRS, INSERM, INRA, IRD, CEA, etc.). These institutes would therefore be, among other things, meeting places favorable to collaborations between the private sector and public researchers, with the possibility for each to have access to the most advanced research of the moment.
Villani report, which research topics?
The report recommends a balance between the dominant themes of the moment (machine learning, data science, big data) that need to be further developed, and those that are less publicized today (semantic web, distributed AI, etc). The interest of this balancing act is mainly to be able to develop fundamental research in areas neglected by the GAFAs in order to be able to differentiate and take market share in an anticipated manner. More generally, it is essentially a matter of encouraging interdisciplinary research, with 3IAs specialized by theme, bringing together researchers and welcoming professionals from different backgrounds.
Bringing together researchers, students and companies
According to the report, there is a clear need for companies to recruit and retain high-level engineers capable of developing clean technology solutions. The raison d'être of the 3IAs would therefore be to meet this need, by training students with a combination of fundamental research and practical solutions. Whether through formal events or regular participation of companies in one of the institutes, they could maintain a sharp technological watch, exchange views and advice, as well as quickly implement joint projects. The success of these institutes depends on the administrative facilities (decision-making circuit, agreements on intellectual property, etc.) that would be granted to companies to participate in this collaboration.
The status granted to companies (permanent, affiliated, temporary members, etc.) within these institutes would be graduated according to their financial involvement in the operation of the 3IA. This status would determine whether or not they participate in seminars, joint research, and whether the institutes welcome in-house researchers from companies. This interaction between companies and research can potentially boost the innovation movement and emulation linked to artificial intelligence. The optimization of available data in the context of Big Data is a major economic challenge today.
A very large amount of data is available to companies since the implementation of open data of public data and the opening of certain private data, but this mass of information can only be really valued if this data is analyzed and made meaningful. Today, only techniques based on artificial intelligence, such as machine learning, are likely to make this data speak and optimize its economic value.
Deepening these techniques and making them operational in companies is the challenge of this proposal in the Villani report, which aims to get public research and the business world to work together.
*The four giants of the Web: Google, Apple, Facebook, Amazon
