Sites Inria

Version française

Big Data and Artificial Intelligence

Carole Céard - 27/10/2017

For Nozha Boujemaa, these technologies are at the crossroads between the digital sciences and the human and social sciences

Nozha Boujemaa - © Inria / Photo S. Tetu - La Company

This year, the European Commissionand the BDVA (Big Data Value Association) are joining forces in order to organise the "European Big Data Value Forum" (EBDVF) which will take place at the Palais des Congrès in Versailles from 21 to 23 November 2017. Inria is a partner and co-organiser of this European event dedicated to Big Data. This first edition will bring together industry professionals, researchers and political decision-makers to discuss the challenges and opportunities of the European data economy and innovation through data in Europe.

Meeting with Nozha Boujemaa, research director and director of the Institut DATAIA (Data Sciences, Intelligence and Society Institute), to find out more about this event and the challenges and issues of Big Data and artificial intelligence technologies in France and in Europe.

> Inria is strongly involved in the organisation of this first edition of the EBDVF. What exactly is involved?

Nozha Boujemaa: The EBDVF proposes themed conferences and presentations that will focus on advanced industrial applications of the Big Data technologies, innovative case studies of the data economy, inspiring visions of the future and reflections on European policy and the financing of R&D&I[1] in this field.

This edition focuses on the theme of "Trusted AI in Smart Industry" . Data are now omnipresent in personal, professional and industrial contexts. More and more digital services call on algorithmic decisions in numerous fields, from the processing of personal data to the sensitive data in critical industrial embedded systems . The transparency and responsibility of these data-based algorithmic systems take on an increasing importance for trust in digital services in B2C[2] and B2B[3] contexts. They constitute factors of competitiveness for data-led innovation.

During this event, we will have the opportunity to exchange ideas with a multitude of professionals in charge of digital transformation stemming from the world of industry, like Airbus, Vodafone, Siemens, Thales and many others, from the academic world with researchers from the Max Planck Institute, Inria, Sciences Po but also institutions such as the European Commission, for example.


> Artificial intelligence (AI) and Big Data are two fast-growing technologies whose convergence offers new perspectives. What are the challenges of such an alliance?

Nozha Boujemaa : In reality, the separation of the two is totally artificial - literally! There is a strong duality between data and algorithms which gives them meaning and allows knowledge and support with decision-making to be derived from them. Data without algorithms to process their content represent an inert mass, of no value at all . The innovative services stemming from artificial intelligence are based on data that have been used to train, test and design them. An algorithm - regardless of its performance - can only provide meaningful results if it has data appropriate for the problem to be solved at its disposal. Data are the vital fuel of AI algorithms .

Today, although learning algorithms are, for the most part, shared in open source, it is noticeable that, on the other hand, this is only rarely and/or with difficulty the case with data.


> What are the challenges that Inria is addressing in the field of Big Data and artificial intelligence?

Nozha Boujemaa : The fields we are talking about pose numerous challenges of different types. This year, Inria has worked on the elaboration of its scientific strategic plan. In it there are numerous challenges relating to the data sciences and to the different areas of application with a strong technological and social impact. 

I will focus on just one example of the challenges here, so this is far from exhaustive: the development of ethical and responsible-by-construction methods to enable the provision of artificial intelligence services that would be trusted by their users, be they citizens or professionals.

Indeed, this is the subject I have chosen for the EBDVF this year, as I am convinced of the critical aspect of digital trust on the rolling-out of the digital economy. The transparency and accountability of the algorithmic systems represent challenges at the crossroads between the digital sciences but also the human and social sciences. 


[1] Research & Development & Innovation

[2] Business to consumer

[3] Business to business


Keywords: Nozha Boujemaa Inria Saclay - Île-de-France research centre Artificial intelligence Algorithms Big data