Criteo and Inria join forces to develop trusted AI Computerworld

Inria announces a strategic partnership with Criteo to develop trusted AI through the pooling of teams. The advertising retargeting company wishes in particular to study the impact of artificial intelligence in advertising content.

Criteo, specializing in online advertising retargeting, announces a research partnership between the Criteo AI Lab and Inria (the National Institute for Research in Digital Science and Technology). Lasting five years, this agreement aims to develop “ethical and trustworthy” AI. Founded in Paris in 2005, Criteo has been listed on the Nasdaq since October 2013. The company, which specializes in business data analysis, has launched an AI research laboratory composed of 60 engineers and 30 researchers, in order to support research in this domain. For Inria, working with Criteo is part of the framework, initiated in 2018, of strengthening strategic agreements with French companies. After Naval Group, EDF and TotalEnergie, Criteo is the 4th company with which the research organization has created a joint project-team.

Concretely, this partnership will cover all future collaborations between the two parties on artificial intelligence with the creation of a joint team, called Fairplay and led by Vianney Perchet, principal researcher at Criteo, at CREST and professor at the ENSAE as well as Patrick Loiseau, Inria researcher. This team will be made up of five academic researchers (Inria, ENSAE and Ecole Polytechnique) and four experts and engineers from Criteo.

Responding to different challenges with AI

At the heart of this collaboration, the two organizations wish in particular to study the impact of AI on the design of so-called “transparent and fair” marketplaces for users. Clearly, the objective is, in the long term, to better control the algorithms and data processing chains, and to feed those of an ethical and trustworthy AI. Through its research, the Fairplay team wants to answer the following issues: how to provide access to relevant and personalized content, while taking into account the imperatives of the protection of personal data, and in particular in certain cases, the lesser knowledge of users and their uses? How to ensure that an advertisement or a job offer can be seen in a non-discriminatory way?

According to the standards of research in artificial intelligence, the work of the team will be communicated in scientific journals and through conferences, and the code developed will be shared in open source. “As coordinator of the research component of the National Artificial Intelligence Plan, Inria is committed to accelerating France’s leadership in responsible AI,” said Jean-Frédéric Gerbeau, Deputy Chief Executive Officer for Science at Inria.

Promote R&D between public and private

This collaboration between the private and the public will also give Criteo the opportunity to continue its involvement in scientific research, responding to industrial issues. The company therefore plans to develop ethical data management tools and make them accessible to the research community and companies. “Criteo continues, with this partnership, its policy of open research and wants to take advantage of this collaboration to address crucial questions raised by AI, such as those of trust, fairness and confidentiality,” said Romain Lerallut, VP Engineering and Director of the Criteo AI Lab. It should be noted that the first results of the work will be communicated within a year.

We would love to thank the author of this short article for this outstanding material

Criteo and Inria join forces to develop trusted AI Computerworld

Check out our social media profiles and also other related pages