Astrocytes have long been neglected in the field of Neuroscience. Now that their multiple roles in brain physiology are recognized, the next challenge is to understand their implication in disease. Astrocytes become reactive in virtually all pathological situations affecting the brain. This response is defined by conserved morphological changes but its functional impact on brain homeostasis and neuronal survival remains debated, especially in the context of neurodegenerative diseases. Such conflicting results are classically justified by differences in molecular triggers or experimental conditions, but instead, they reveal that astrocyte reactivity is a much complex response than initially thought. It is now time to address this complexity and to decipher the molecular mechanisms governing the different states of astrocyte reactivity. If we are able to identify these mechanisms then, we will be able to target them for therapeutic purposes. The DecodAstro project aims at breaking the code of reactivity, thanks to cell-specific, sensitive and novel techniques applied to relevant in vivo models of neurodegenerative diseases. We recently established the importance of the JAK2/STAT3 pathway for reactivity in neurodegenerative diseases. Here, we will identify additional master regulators of astrocyte reactivity in an unbiased manner, through genome-wide screening in vivo. In addition, we will monitor the heterogeneity of astrocyte reactivity with time-lapse two-photon microscopy in situ and determine how specific signaling networks translate into unique transcriptional programs and functional outcomes. The paradigm shift is to no longer consider averaged populations of reactive astrocytes with inevitable discrepancy between experimental conditions, but instead reveal and study cellular heterogeneity of reactive astrocytes in situ. DecodAstro will significantly improve our understanding of brain response to disease. By addressing the emerging theme of cell-to-cell variability in biological responses, this project will open new conceptual and technical avenues in Neuroscience and adjacent fields.

AI4LEA project aims at providing Technological Autonomy to Law Enforcement Agencies (LEAs) to prevent and fight against crime and terrorism. The project will create a long-lasting community of LEAs and the R&D industry to foster effective collaboration. Within this community, AI4LEA will explore the use of different tools and techniques, in order to define, develop, share, and evolve open source Artificial Intelligence (AI) based technology solutions for LEAs. AI4LEA adopts a co-development strategy based on fluid, frequent, and fruitful collaboration between all stakeholders, including short development cycles and face-to-face “Hackathons” every 6 months. After the hackathons, LEAs will be able to deploy and test the tools in their own premises and with their own data, while providing feedback to the whole AI4LEA community. The project will expand the initial ASGARD H2020 project by enrolling new member states (essentially by involving new LEAs), by addressing new use cases organized in operational areas (cyber-crime, border security, …) and by disseminating the results through European agencies such as EUROPOL. The definition and design of the AI based solutions will rely on (1) forensics, intelligence and foresight processes, (2) use cases and scenarios driven by end-user needs, (3) Service Oriented Architecture (SoA) technologies and achievable challenges, and (4) Social, Ethical, Legal, and Privacy aspects (SELP). AI4LEA will develop interoperable and easy-to-use sets of tools complementing LEAs’ current systems. The project will have a particular focus on technological breakthrough by exploring the use of AI towards security at and beyond the state-of-the-art. Innovative solutions will help LEAs investigate traditional forms of crimes where digital content plays a key role, cyber-dependent and cyber-enabled crimes. AI4LEA will perform research into the capabilities and weaknesses of AI systems to prevent their malicious usage by cyber criminals. At SELP and privacy dimension, the project will be fully compliant with new legislative framework resulting from GDPR EU directive with emphasis placed on privacy by design and on societal impact. By the end of the project, AI4LEA will extend an active and sustainable community of practitioners at the EU scale with a valid and sustainable model for all participants with successfully delivered and evaluated tools & a collaborative infrastructure allowing the sharing of knowledge while protecting sensitive and sovereign data. Project’s main impact will be to enhance LEAs’ autonomy, efficiency and capabilities in forensics, intelligence, and anticipation/foresight by delivering a set of easily configurable and deployable tools and applications. The tools to be delivered will be prioritised by LEAs. Each tool will be designed and developed to tackle a specific task, whether it be data or task driven. User-friendly and easy-to-use applications will allow very quick definition and set up of ad hoc data acquisition, processing, analysis, and exploitation workflows to tackle the specific needs of each investigation. AI4LEA will also improve LEAs’ capabilities for trans-border LEAs knowledge and data-exchange and collaboration. AI4LEA implements agile and modern continuous development together with integration methodologies and short development cycles. This process ensures the LEAs in the project have early and frequent access to the project results so that they can provide prompt feedback to re-prioritise the work plan if needed. The project adopts a streamlined management and coordination with few partners per task and few tasks per partner. An efficient governance and decision-making mechanisms is set to simplify management and coordination and to facilitate prompt and appropriate issue/conflict resolution.

The main objective of the project is to develop a bipolar battery by using safe materials in compliance with environmental regulations. Because of the performances achieved so far, the high cyclability the Ni-Zn battery technology developed by SCPS was chosen. The bipolar Ni-Zn battery can meet the required specifications to power a robot, while offering enough design flexibility to meet other constraints.

The MOSARIS Common Laboratory- Methods and Tools for Risk Analysis of Smart-Grids) aims to build an innovative environment dedicated to assistance for conducting security risk analysis in the domain of industrial networks like “Smart-Grids”. The proposed program deals with security of cyber-physical systems and critical infrastructures, such systems which include both classical industrial infrastructure and NTIC (New technologies of Information and communication). Considering the high potential of development of such systems in the near future as well as the implied challenging issues both in the economical, strategic and global security; it is necessary to propose new analysis and securisation approaches adapted to this context and supported by efficient tools. The analysis culture was until now conducted according to different manners into two worlds: the world of industrial systems and the world of information systems. It is attested that the introduction of NTIC within industrial infrastructures raises new risks and increases their vulnerabilities. Security analyses must thus be adapted for this new context in order to get a better control on the overall aspects of the vulnerabilities inherent to such systems including possible interactions between security goals and other criteria, as well as the follow up in the time and accompaniment of system evolutions. Conducting such studies was until now largely supported by human expertise; this will soon no longer be possible. It becomes necessary to support security analysts with new methodologies supported by tooling devoted to this new context. The objective of the collaboration is to propose methods and the realization of a software environment dedicated to such problem and compatible with standards like ISO/IEC 27005 or guidelines such as the Ebios methodology, improved to take into account the specificities of this new context. This tool will be designed to be flexible and progressively enriched. This realization will be made possible by the cooperation of two expertises: • The expertise of CoESSI a consulting company in security of information systems that has also a great competence in the domain of Smart-Grids. • The expertise of the Laboratoire d’Ingénierie des Systèmes Embarqués (LISE from CEA), au sein du CEA) in the domain of system modeling and analysis. In particular the Laboratory has experience in multi-criteria evaluation and safety analysis of embedded systems. The objective is to enrich the methodological approaches developed by the tow partners and to design and develop software tools that will permit to CoESSI to develop its activity in the domain through an automation and improvement of current practices in systems security analysis. These tools will rely on the open-source system modeling framework developed within the laboratory: Papyrus (www.eclipse.org/papyrus). They will permit to conduct security risks analysis from system models and dedicated knowledge bases coming from the expertise of security analysis experts in the project. In a first stage, a first methodological approach inspired from the current practice of CoESSI will be implemented and experimented on case studies. The creation of the laboratory intends to provide a cross-fertilization locus to improve progressively the methodologies supported by the tool. In the future more dynamic strategies will be supported. Tools developed will be exploited by the CoESSI society within its consulting activity and will be commercialized as a complementary support for its clients.