Inria
a hive of goodwill

Maureen Clerc, could you outline the main areas of research at Inria Côte d’Azur?

First and foremost, we work on data science, using models that can represent transport networks, cellular communication networks in biology, or medical knowledge.

We are trying to improve knowledge in a variety of fields using methodologies from the digital world. Specifically, we link concepts and make data speak to us in order to better understand complex events.

We also conduct more fundamental research with researchers working on computer forensics or communication networks using software virtualisation, for example, or post-quantum cryptography.

AI is playing an increasingly important role. Some teams focus on artificial intelligence as their main research area, while others do not—but are nonetheless making growing use of AI in their work.

Inria is heavily involved in the 3IA Côte d’Azur Institute. What does that involvement look like in practice?

Many of our researchers hold 3IA chairs. This means they have been selected by an international panel, which deemed them qualified to implement or develop new AI techniques. These researchers are also engaged in training PhD students in AI and supervising postdoctoral researchers who contribute to the work of our teams.

In fact, a real dynamic has been created on the site, where AI researchers from different fields come together to share their knowledge. In addition, our 3IA chair holders are involved in teaching and contribute to industrial collaborations, which have been strengthened since the creation of 3IA Côte d'Azur. 

The Scientific Director of 3IA Côte d'Azur, Nicholas Ayache, is a renowned researcher at Inria in the field of digital health. To what extent does AI contribute to the development of particularly effective analysis and diagnostic tools?

With its ability to quickly process large amounts of data, AI provides doctors with assistance that they will not be able to do without in the future. It can also re-examine cohorts in ways that humans cannot.

Inria Côte d’Azur is also participating in the adventure of a new IHU in Nice, called RespirERA. AI is one of the strengths of this university hospital institute. In Sophia Antipolis, we have leading teams that have long been interested in how to automate the processing of medical images, as well as all other information accompanying patient medical records.

SustAInability at Inria

You are also working extensively on smart territories. To what extent can the optimisation of urban infrastructure and transport systems play a role in reducing carbon emissions?

At our centre, we have a researcher, Paola Goatin, who is working on road traffic regulation. Using partial differential equations, she is looking for the optimal speed to limit congestion, particularly at motorway entrances and exits. It is clear that having much smoother traffic has an effect on reducing carbon emissions. This is also of interest to citizens in their daily concerns. We are a little frustrated that we have not been able to collaborate with motorway companies to set up real experiments in France. However, Paola Goatin is working with colleagues in Berkeley, California, where real experiments are being carried out.

In the transport sector, we collaborate with Instant System, a company based in Sceaux that has major contracts in the Paris region with Transdev. Our team, which is responsible for optimising networks, is assisting Instant System in implementing effective strategies that combine carpooling and public transport, thereby promoting multimodality. I firmly believe that this solution will help reduce carbon emissions.

While AI enables progress in environmental matters, its widespread use remains very energy-intensive and contributes to global warming. What is Inria's position on this?

There are structural and organisational aspects. In terms of calculations, we do them either locally or nationally, on platforms that have been designed to be as efficient as possible.

Here, locally, we have a responsibility to ensure that our data centre is as environmentally friendly as possible. Currently, its utilisation rate is around 60%. We can try to use it even more, because once equipment has been built, it is better to use it as much as possible.

We're also working on how to cool it. A data centre generates a lot of heat and keeping it constantly cool requires a significant amount of energy. We have a plan to completely refurbish our current data centre, with a focus on improving its energy efficiency. We're also looking into geothermal energy as a way to cut down on our energy consumption. Our researchers are also working on ways to optimise electricity use for networks and computing. Their solutions may be implemented in the future to better schedule computations and the entire process in order to reduce energy consumption.

Should we change the image and use of AI?

Undoubtedly, and we are not always obliged to use AI. Now that Chat GPT exists, people are turning away from traditional browsers. I believe there is a lot of education to be done on the energy cost of different solutions.

Together with CNRS and Université Côte d’Azur, we launched Terra Numerica, which now has a dedicated space in Sophia Antipolis. It’s a place where people of all ages can gain a better understanding of what lies behind our screens—by exploring digital concepts and objects in a hands-on, playful and accessible way.

While there is much to be done to educate the general public, the same is true for researchers. To reduce our footprint, we must be reasonable in the way we conduct research. This conflicts with the pursuit of innovation. It is a delicate balance that is not easy to manage. We must restrict ourselves to reasonable proportions.

Making AI more environmentally friendly also requires interdisciplinarity. What do you think?

Interdisciplinarity allows us to get the most out of a corpus, because in order to manage data flows in a relevant way, we need to be able to exploit them. The challenge here is to extract all the information we may need for the future without keeping data that is of no use.

One strategy adopted by the Montpellier branch of Inria Côte d'Azur has been to set up an interdisciplinary team bringing together computer scientists, ecologists, agronomists and climatologists. This team (Iroko) is involved in numerous European projects in which it examines a large amount of data from a variety of countries.

Is there any point you would like to emphasise?

One point we haven't discussed concerning the climate, which is important for our site. It concerns environmental risks, one of the strategic topics at Université Côte d'Azur. We are already in a seismic region, but on top of that we have the sea and all these mountains, which occasionally cause disasters during events that are unfortunately becoming less and less rare.

Inria can play a role in helping people on the ground to make planning decisions. We have talked about road traffic, but this can also apply to coastal planning. In Montpellier, we have a team working on coastal flow modelling and urban flood modelling.

We also have a team here that creates environments that look real but are based on simulations and artists' creations. Creating scenes as they would be if floods, avalanches or landslides occurred can make planners aware of what needs to be done in these situations. To help with decision-making, we have a pioneering role to play in being able to simulate and predict, among many possibilities, what could happen.

Terra Numerica is a joint project between the CNRS, Inria and Université Côte d'Azur, whose aim is to give all citizens access to high-quality digital culture resources. It brings together a diverse range of partners, including the French Ministry of Education, and has developed an original, attractive and unique approach to scientific mediation, dissemination, sharing and networking between digital stakeholders. The project focuses on developing interconnected networks of animated digital culture spaces and offers fun resources for all audiences with a clear focus on promoting hands-on learning and interactivity. This includes the Terra Numerica @Sophia totem site, a network of partner spaces in the South Region, and many outdoor activities as well as online activities. Terra Numerica deploys tools adapted to all levels, from discovery to training, to address the fundamentals, applications, challenges, IT heritage and digital professions. More than 120,000 people have been educated or trained in six years.