SPIE and INSA Lyon create an “Artificial intelligence for behavioural analysis of streams in digital infrastructures” chair

Malakoff, 13 June 2022 – SPIE ICS, the French digital services subsidiary of the SPIE group, the independent European leader in multi-technical services in the areas of energy and communications, is creating an academic research chair that will investigate issues related to the analysis and smart processing of data in communication infrastructures, as closely as possible to sources of production (applications, users, networks, businesses). SPIE ICS is creating this chair in partnership with INSA Lyon (National Institute for Applied Sciences of Lyon).

This chair aims to develop research in order to make digital infrastructures more efficient and contribute to a responsible digital transition. Over the next five years, an AI-assisted search engine will be developed that is capable of using data close to their source of production in real time and locally processing that data in a responsible way.

A move towards managing data more responsibly 

Data management is a catalyst for transforming organisations. It represents a key pillar of SPIE ICS’ 2021-2025 Strategic Plan.
In turn, INSA Lyon hopes to consolidate its position as a groundbreaking academic force when it comes to supreme, sustainable processing and communication systems that have a lower carbon footprint. 

Nowadays, the development of “smart” applications (IoT, Smart Cities, Smart Buildings, Smart Factories, Smart Grids, etc.) produces a significant amount of data. Analysing this data in transit creates value and produces new use cases at local level: predictive behavioural analysis, dynamic “scaling” of resources and infrastructures, reduced costs, prevention of cyber threats or real-time traffic auditing. SPIE ICS hopes to boost its data-oriented approach with the objective of creating operational value and developing local services by utilising this scientific partnership which falls under the chair programme sponsored by the INSA Lyon Foundation.

Identifying new breakthrough technologies

The work developed as a result of this chair will make it possible to create an AI engine using edge computing  which is capable of more effectively understanding the data streams that pass through infrastructures, thereby making them easier to see, understand and process. The research will concentrate on the technologies used to analyse and process local data. It will be coordinated by chairholder and CITI lab director Frédéric Le Mouël in conjunction with the CITI and LIRIS laboratory teams. 

The research will focus on the various breakthrough technologies: 
-    Data-oriented infrastructures and protocols: the challenge here is that of determining how to extract information in transit in order to aggregate useful knowledge while reducing its energy footprint
-    Embedded artificial intelligence: to understand behaviour in an uncontrolled environment, for example, with an AI engine embedded into low-power systems and working on partial data, with collaborative and reinforced learning techniques  
-    Geo-distribution of resources and services: this involves devising auto-managed infrastructures (without human intervention) that guarantee operational compatibility of services and systems whereby these can be aligned with a data-oriented approach.  

Digital trust

“Digital trust is integral if organisations want to control their data and services”, explains Karen Luzignant, the Business Director for Grand-Est at SPIE ICS. “The use of proprietary equipment opens up new services but also results in organisations becoming dependent on companies that are subject to different laws. It is essential to keep infrastructures and data sovereignty under control given that data and data-related services are a key source of profitability.” 

Marie-Christine Baietto, the Director of Research and Development at INSA Lyon, believes that “the ‘digital transformation’ of society (‘the digital society’) presents a major challenge for information science and technology operators and calls for increasingly fruitful partnerships with human and social sciences. This transformation is based on new information modelling, capturing, processing and communication processes. Big data are collected and now represent a raw material to be converted. The resurgence of artificial intelligence, based on greater computing power and powerful automatic learning techniques, is resulting in increasingly robust and efficient solutions for understanding, analysing, devising and deciding upon new services in every sector of the economy, while at the same time addressing decisive lines of research into energy consumption, security, reliability and agility.  These are all responses to what our partners and society require.”