December 8, 2016 | IT security: software secrets revealed
The BINSEC platform analyzes executable software code for enhanced IT security. The platform's tools are derived from formal critical-code analysis methods.
Software security is one of today’s major concerns. And yet, current security analysis methods focus mainly on software source code, which cannot always be accessed. “Even when you can get to the source code, it can be challenging to analyze it thoroughly enough to detect malware or bugs introduced by compilers,” said a researcher at CEA Tech institute List.*
Researchers at List drew on their broad, deep knowledge of IT security to come up with reliable analysis tools to fill the current gaps. They started with formal methods—state-of-the-art source-code analysis techniques used for critical systems—and modified them so that they could be used to analyze executable code. “Formal methods are sensitive to the program’s behavior rather than to how the program is coded. These methods are also capable of exploring all, or almost all, behaviors.” The tools List developed are more robust and reliable than the syntax analysis used by most software manufacturers. This type of analysis provides a cursory examination of the executable code or analyzes a random selection of certain behaviors.
The tools developed by List are the result of R&D partnerships, and have been packaged together to form the BINSEC open-source binary code analysis platform. The platform was tested successfully and will be unveiled at the Black Hat Europe trade fair at the end of this year.
* Source code is written in human language by programmers; the compiler is a program that translates human code into executable code (machine language).
November 22, 2016 | A discreet assisted-living safety solution
Schneider Electric's future home surveillance system will be non-invasive and personalized for optimal efficiency. List, a CEA Tech institute, developed the smart capabilities that power the system.
Schneider Electric is currently developing an affordable, non-invasive assisted-living solution to detect unusual situations in elderly people's homes. And the solution leverages two major innovations. First of all, it uses a low-cost network of sensors selected because they are both non-invasive and flexible.
The sensors are not worn by the patient; they are installed in the living environment and provide information about the occurrence of events like opening and closing doors, the patient's presence in strategic rooms, and more.
Plus, the fact that the sensor network is personalized makes it even more effective. "We developed an algorithm that factors in the patient's habits and time of day to pinpoint any abnormal situations."The algorithm just needs a simple data log to automatically determine "reasonable" alert thresholds, ensuring an excellent compromise between reducing false alarms and achieving a high incident-detection rate. The statistical method used works independently of the type of sensor and the algorithm improves its performance over time. The solution has been evaluated in test homes and is currently being assessed by the manufacturer.
17 novembre 2016 | Rencontres du Numérique de l'ANR : le List décroche le Prix de l’Impact Economique
Les 16 et 17 novembre s’est tenue la 2ème édition des Rencontres du Numérique de l’ANR à la Cité des Sciences et de l’Industrie à Paris. Parmi les 200 projets de recherche financés par l’ANR en compétition, c’est Florian Gosselin, coordonnateur du projet MANDARIN, qui a obtenu le Prix de l’Impact Economique*.
Le projet collaboratif MANDARIN (Haption, INRIA, Université Technologique de Compiègne et Renault), avait pour ambition le développement d’un dispositif haptique permettant une manipulation dextre, naturelle et intuitive dans les environnements de réalité virtuelle.
Les chercheurs du List ont développé un gant haptique à retour d’effort se distinguant des solutions existantes par sa légèreté, sa compacité, ses performances et sa facilité d’utilisation par des non-initiés.
Une version pré-industrielle est actuellement en cours de développement par la société Haption, startup du List.
*Le projet MANDARIN est arrivé ex-aequo avec le projet TRIMARAN
Pour en savoir plus :
- Rencontres du Numérique de l'ANR
- Présentation du projet Mandarin
- Actualité du 10 novembre 2016 « Un gant à retour d’effort pour une manipulation virtuelle précise »
November 10, 2016 | Force-feedback glove for precision virtual handling
CEA Tech Institute List presented its innovative virtual-reality system at the Innorobo trade show in May 2016. The compact, lightweight glove lets users efficiently and accurately work with items in their environment.
A new advance in haptic force feedback systems has been made by List, the result of R&D conducted under French National Research Agency project Mandarin. These systems are used to remotely-control manufacturing robots and to let users interact with digital models in virtual-reality environments. List's haptic force feedback glove is more compact and offers a higher level of performance than other current solutions.
The particularly compact, lightweight prototype packs in several patented innovations. Typically, each finger is powered by three motors to either provide resistance to or to guide the opening and closing of the hand; the List prototype uses a single force feedback motor. The prototype also leverages a mechanism to simulate finger-pad deformation, ensuring 3D haptic feedback.
The researchers also developed a compact gear system to make the actuators more compact, lighter in weight, and reversible—a much more "transparent" solution than what was previously available. Optical sensors placed around the axis of each joint were chosen to improve overall system integration.
List presented the product to the public at the last Innorobo trade show in Paris; a prototype a step closer to industrial scale-up is expected to be completed by end-2016.
November 4, 2016 | Social media: savvy users are safe users
List, a CEA Tech institute, recently participated in EU project USEMP, developing new tools to protect users' personal information on social networks. The DataBait platform won the Best Project Award at ICT 2015*.
If you are like most people, you probably worry about privacy when you post a photo or comment on social media. Soon, DataBait could do the worrying for you by analyzing the risks before new content is published. List developed the new software as part of EU project USEMP to enhance the protection of personal information online.
IT specialists, sociologists, and legal experts collaborated on this multidisciplinary project to bring users profiling tools as powerful as those used by the social networks themselves. "The idea is to make sure that users are aware of the information that can be mined from what they post on social media, and how their information can potentially be used."
List contributed to the project by developing image recognition and analysis tools and text mining technology. The former leverage the use of neural networks (an area in which List has built up renowned expertise); the latter is based on sophisticated comparative processing techniques.
The first version of the software is currently undergoing validation testing. List is seeking partners to take the tool beyond the EU project that facilitated its development.
*Conference organized by the European Commission in Lisbon on October 20–22, 2015.