According to Mathieu, AI at Scalian is anything but complicated!

  • Expertise

    25 March 2022

Intelligence Artificielle

Artificial Intelligence is a term or concept that raises a lot of questions, often because it is associated with fantasy movie stories and portrayals. Mathieu Damour, Technical Manager, tells us more about AI and explains the reality of Scalian’s expertise in this field, which is anything but artificial!

Can you briefly summarise the history of AI?

Artificial intelligence, or AI, was originally a fuzzy concept from the 1950s. It was a fantasy, often movie-based, vision of an intelligence capable of mimicking biological thought processes, thereby reproducing this human or animal capability and being able to learn and adapt to its environment. Since 2010, technological innovations in computer science and mathematics have allowed us to re-analyse the decades-old concept of neural networks, a type of data processing based on biological thinking, rather than just mimicking it.

The world is now in a position to develop incredible innovations and reach amazing objectives: the fantasy in Spielberg’s films is no longer fantasy, it’s a reality! It is not a reality that comes at the expense of humans, who get replaced by AI. The aim of AI is to help humans, to make them more productive and allow them to devote their time to high value-added activities, not to replace them. An example is in the medical sector, where cancerous cells can be detected even before a doctor can see them!

What do we actually mean by AI at Scalian?

Scalian’s DNA means the company is positioned in the aeronautics and space sector. For example, the AI at Scalian focuses on improving the energy efficiency of aircraft by optimising the development of navigation systems. We also work on problems linked to detecting objects or living creatures in pictures, for different purposes. Detecting anomalies is also a key issue on a manufacturing level, because it reinforces the quality of our partners’ products and ensures the reliability of their production line. For our clients, AI is a vector of absolute confidence. To provide them with comprehensive support, we offer training courses and awareness-raising services. Together, we build their roadmap to transform their daily lives, by exploiting the value of their data and turning them into useful tools for their activity.

****Scalian’s vision of AI: responsible, equitable and tailor-made****

Scalian is part of the DEEL project, involving a consortium of partners and industrial companies that work together on a national and international level to make artificial intelligence more responsible, more understandable and more equitable for all of society. At the same time as the resurgence of AI, there was a boom in the field of Big Data – a real gold mine for the data science field, but which consumes a lot of energy to store data that are sometimes not very usable or valuable. Our responsibility is to help our clients upstream of their data acquisition processes and guide them towards the useful part of their data and thus help them use fewer of their resources. In short, more responsible, smarter data consumption! For several years now, cultural, human and technological barriers have been overcome one by one, but we still need to find powerful hardware resources that break down the budgetary barriers that make AI seem broadly inaccessible. What we are able to do is adapt our solutions to our clients’ resources; we take a pragmatic view, based on our clients’ business reality. We provide our clients with support through training, raising their awareness of the issues and developing pragmatic, tailor-made models for them that meet their exact needs (no more, no less). Our approach is clear.

The AI teams at Scalian

Personally, I have the pleasure of managing a team of experts in Data Science and Deep Learning, from the Bordeaux office. We work alongside other specialists who look after integrating this expertise into our clients’ systems. This actual deployment aspect is extremely important because, while it is becoming increasingly easy to develop this type of algorithm, it is not so simple to deploy them and industrialise them. We should note that Scalian employs a large number of data scientists, but also has many employees in our various entities around France and internationally specialising in other professions that gravitate around this sector. Here, I’m particularly thinking about the Lab teams and our CEN Simu staff, as well as our colleagues in Sophia Antipolis and Indizen in Spain.

Which kind of profiles can join these teams?

Coding experts, preferably in Python. Beyond that, specialists in different kinds of libraries that are used to handle and mine the data, or to build and train Machine Learning and Deep Learning models. A candidate’s main quality needs to be his/ her curiosity. You need to enjoy exploring the latest innovations concerning algorithms and iterative testing of models. Our core business is to industrialise our leading specialities, but our goal will always be to push back the limits of what is possible!

AI