Data and knowledge analysis

Military intelligence turns to gaming for innovation

Date:
Changed on 10/05/2022
How can you present your work and your innovation needs to researchers when your work involves handling classified defence information? This was the challenge faced by the French Directorate of Military Intelligence, until Inria came up with a unique solution: researchers could carry out the work of analysts using a simulation, Battle Lab Rens, and test out innovative new ideas on a specialist platform, Intel Lab.
Informations pour le briefing
@Inria / Photo B.Fourrier

First full-scale test of Battle Lab Rens in late 2019

Whether it’s to support troops carrying out operations or for strategic monitoring, the Ministry of the Armed Forces requires accurate and reliable information on potential enemies’ operational resources and their intentions. This is the job of the Directorate of Military Intelligence, a body which is not particularly well-known to the wider public, despite employing more than 2,000 individuals.

By definition it handles information which is inaccessible to anyone without military security clearance,including industrial or scientific partners. This is what prompted it to design Battle Lab Rens (BLR), a simulation which was first tested in 2019.

“We brought together eight individuals from industry, research and the French defence procurement agency”, explains Thierry Assonion, head of the BLR project. “They played out a fictional scenario inspired by the Libyan crisis in 2011, employing real data from the field which was not covered by military secrecy.” 

Six hours in the shoes of an intelligence analyst

The players spent six hours doing the job of an analyst: reading overviews, receiving intelligence reports from multiple sources (intercepted messages, aerial images, internet, human intelligence, etc.), drafting memos and status reports as requested by the political/military authorities, etc., supervised by experts from the field.

“They really enjoyed the immersive experience, which was much more vivid and illustrative than any presentation could have been”, recalls Thierry Assonion. “BLR showed that it could help our partners in research to understand our problems, giving them ideas for innovations.” 

Processing big data quickly and efficiently

In military intelligence, innovation is currently focused on IT equipment. This is because the main issue is information overload: sources or “sensors” discharge flows of text, images and videos, which analysts can’t process without digital information processing tools.

“We are required to sort, handle and interpret large quantities of heterogeneous data, becoming quicker and more relevant all the time”, explains Thierry Assonion. “This data then has to be reproduced in an intuitive format in order to inform decision-making.” 

Verbatim

We operate on two levels”. “The first involves improving Battle Lab Rens and developing Intel Lab, a platform dealing exclusively with research and experimentation. The second is actual innovation: by participating in these gaming sessions, scientists get first-hand experience of intelligence challenges, enabling them to come up with solutions.

Auteur

Frederique Segond

Poste

Head of mission Securite-Defence

A more complete version of Battle Lab Rens

Inria’s Security and Defence Mission currently has six permanent employees and is establishing ties with nine regional Inria centres in order to identify points of contact. While this has been happening, it has also got to work on its tools.

The 2019 version of Battle Lab Rens, based on the Microsoft Office suite, has evolved into a more sophisticated web application. It allows users to sort information from “sensors” into dedicated spaces, to implement scenarios, to incorporate maps and aerial photographs, to manage the profiles of participants (coordinators, players, experts, etc.), to provide resources on the military intelligence process, etc.

“It’s a remarkable tool”, says Thierry Assonion, who has continued to produce other scenarios inspired by past crises. “Our innovation partners get the chance to spend a day in the shoes of an analyst, which is unimaginable in real life for reasons of military secrecy.”

Verbatim

C’est un outil exceptionnel, qui continue à produire d’autres scénarios inspirés de crises passées. Nos partenaires innovation se retrouvent dans la peau des analystes pendant une journée. Un scénario inimaginable dans la vraie vie, pour des raisons de secret-défense. 

Auteur

Thierry Assonion

Poste

Chef du Projet BLR

Intel Lab, a platform to support innovation

Running parallel to this, the Mission has developed Intel Lab, a software and hardware platform with a wider scope. Intel Lab hosts Battle Lab Rens, and in the future will host other applications of the same type: for simulating economic intelligence scenarios, for example, or preventing acts of terrorism or cybercrime. It is also used for research and experimentation.

“Imagine that a brainstorming session following a gaming session leads to a major group launching an R&D project”, says Frédérique Segond. “When the time comes, it must be able to test it in conditions that are representative, and have it evaluated by intelligence experts. Intel Lab makes this a possibility.”

Realistic data for testing developing technology

Here’s an example to illustrate how this works: the surveillance of messages sent via an encrypted social media platform, between members of a group suspected of terrorist activity. Conversations may take place over several months and cover trivial subjects, such as cooking, before suddenly switching to a discussion of a worrying “journey”.

A company or research team developing software for analysing such conversations, detecting shifts and clarifying details of this “journey” - the location, the date, the individuals involved - would not have access to real conversations”, explains Frédérique Segond. “However, they will find these on Intel Lab, where we can generate conversations based on data we have on file.” 

The platform is also opening up to other uses. Over three days it played host to a group of 12 future analysts, introducing them to the fundamentals of rigorous reasoning based on multiple intelligence reports.

At such an early stage, neither Battle Lab Rens nor Intel Lab have given birth to their first innovations.“We are still at the learning stage”, explains Frédérique Segond, “both in terms of the level of detail for scenarios and the format for the brainstorming sessions following the gaming sessions.” But Inria already knows that a number of its fields of expertise will be of assistance to the public body: deep learning, neural networks, heterogeneous data fusion, natural language processing, digital image processing, incident detection, etc. Through its Security and Defence Mission, the institute is standing by ready to deploy.