0M €
EDF-2021-DIGIT-R-FL: Frugal learning for rapid adaptation of AI systems

In times of real-time information availability and exchange, and increasing complexity of situations, artificial intelligence (AI) has become an essential driver for new competitive system solutions. Future military capabilities will include a significant share of systems that will make massive use of AI techniques.

In times of real-time information availability and exchange, and increasing complexity of situations, artificial intelligence (AI) has become an essential driver for new competitive system solutions. Future military capabilities will include a significant share of systems that will make massive use of AI techniques.

Modern AI systems based on Machine Learning and especially Deep Learning techniques usually require many labelled data points to reach acceptable performance. Furthermore, they can suffer from inconsistent behaviours, such as high-confidence failures, or failures in trivial cases. More generally, improving AI systems to take into account new data requires extensive testing by expert developers to avoid regression. These issues severely impact their availability for defence systems, which are characterised by the lack of data, for instance when dealing with enemy intelligence, and by the need for trustable results and rapid adaptation, including from data that cannot be shared with system developers for confidentiality reasons or because of poor connectivity. This is especially important when the information to manage is highly variable or unpredictable and high adaptability is needed.

The challenge is to develop new Artificial Intelligence methods that are able to make use of less training data than current state-of-the-art deep learning algorithms while maintaining similar performance, to provide better control over the output space in order to ensure a more consistent behaviour, and to limit the development efforts when adapting systems to new data. These methods must prove their worth on realistic and challenging use cases representative of military operations.

The aim is to tackle the problem of robustness and frugality in military AI software components to facilitate the development of new systems and their adaptation to the evolution of their environment, including from user supervision, for a reasonable cost, with minimal intervention from expert developers, and without regression. State-of-the-art research on transfer learning, zero- or few-shot learning, active learning, domain adaptation, hybrid AI and other relevant topics should be leveraged to propose new methods to improve AI-based methods, while preserving high performance.

The proposals must cover the following activities as referred in article 10.3 of the EDF Regulation, not excluding possible downstream activities eligible for research actions if deemed useful to reach the objectives:

  • Activities that aim to create, underpin and improve knowledge, products and technologies, including disruptive technologies for defence, which can achieve significant effects in the area of defence.

The proposals must address in particular the following objectives:

  • Design of relevant military use cases where trustworthy and frugal AI algorithms are needed (i.e. targeting specific tasks for which data-greedy algorithms are currently outperforming other methods), and for which representative data can be collected and performance can be measured in an objective way.
  • Development of new methods for reducing the need for data and supervision to train and adapt AI systems (and simultaneous monitoring of the state of the art, which is important in this quickly evolving domain but should take place as a background activity integrated with system development), for example through:
    • simulation and generative models,
    • transfer, semi-supervised, self-supervised, and active learning,
    • hybridisation with user-defined rules.
  • Development of new methods for improving robustness guarantees by design, for example through:
    • new algorithms increasing the robustness of existing neural networks methods that are intrinsically not robust,
    • evaluation and surveillance of both the output space and the environment.
  • Implementation of benchmarking experiments on the use cases to demonstrate the advantages and drawbacks of the proposed methods.
Check the call here
EDF-2021-DIGIT-D-MDOC: Military multi-domain operations cloud

Military operations require higher flexibility and mobility to gain and maintain the initiative. The capability to securely, timely and robustly communicate over all battlespace domains is key for information superiority, mission management and decision support. Therefore, the development of a common shared Information Space with a “Cloud of Clouds” approach, leading to a Multi-Domain Operations Cloud (MDOC), is needed. The ambition is to combine existing and future systems into a federated network and collaborative services in order to enable and support Command and Control for multi-domain warfare. Furthermore, the data collected across domains will open up future opportunities to develop artificial intelligence (AI) enabled solutions for defence.

Information transcends operational domains and multiplies the size of effects in combat. Warfare is no longer segregated into specific domains, as information sharing lies at the heart of cooperation and boundaries of the individual domains are blurring.

A collaborative, efficient and secure information management across land, sea, air, space and cyber domains is key for operational superiority, mission management, decision support, and for future capability development in the area of AI.

An overall military advantage and information superiority will be achieved through complete situational awareness based upon current data from all available sources. In modern warfare, the right information at the right place and at the right time can make the difference in a contested military environment as well as information gaps in the non-real-time reporting chain. Furthermore, if data is collected, stored and made securely available for the purposes of later developing AI solutions, its value would be further maximized.

Currently, information is not adequately shared in military systems and is rather kept enclosed in systems or sub-systems interconnected by local and domain specific interfaces. The lack of information sharing and coordination across all military domains is amplified by the existence of different data models limiting appropriate information exchange and exploitation. This situation may lead to different information interpretation, thus producing multiple situation pictures of the same situation and ultimately allowing taking uncoordinated decisions.

Additionally, the digitization in every domain progressively introduces high-performance systems creating increasing amounts of data that needs to be distributed and shared among various combat actors from tactical to strategic levels. This evolution overwhelms the architecture and networking capabilities of current generation systems and creates a challenge for a new generation operations cloud. Furthermore, the lack of specific rules governing data collection and curation hinders the possible re-use of these data for the training of AI solutions.

The civilian cloud versions use very-high-speed networks for information access and synchronisation and take advantage of decentralized resources in multiple data centres. In the military environment, specific constraints exist (such as high mobility with no reliance on support infrastructures, transmission security and electromagnetic contested environment, limits in networks data rate and availability, limits in local computer and storage resources, disconnected modes, environment and hardening constraints, etc.) that impose a challenge on the direct usage of the civilian solution. In addition, even if many classical IT services (messaging, chat…) are close to their civilian counterparts, operational users request specific applications and services which need to be shared for a real federated multi-domain cooperation (e.g. C2 services, ISR services, tactical situation or logistics, training and exercises).

As commercial cloud concepts and the underlying networking models cannot be applied (or can only partially be used), a specific architecture with cloud technologies has to be set up for military purposes reflecting the needs for special adaptations, especially for the operational and tactical use cases. However, cyber ranges across multiple EU Member States have the relevant infrastructure that would allow them to act as secure data repositories and for training and testing (i.e. sandboxing) AI solutions.

Proposals should address the development of a multi-domain cloud architecture for defence and an associated technological demonstrator, providing a common shared information space and federated services enabling multi-domain operations.

The ambition is to combine data of existing and future systems through a federated network, shared cloud interfaces & implementation of the associated shared services. The ambition is also to make the data securely available for re-use in the promising field of AI.

MDOC must enable and support flexible combined and joint military missions and provide the capability for an accelerated and improved battle rhythm for military operations in a collaborative multi-domain warfare, ensuring adequate level of data protection.

MDOC must include three major components:

  • European virtual or digital platform
  • Catalogue of end-user products and services
  • Tools, interfaces and APIs

The proposals must cover several key aspects and show how they will handle them:

  • The specificity of requirements (operational/technical/environment/etc.) at strategic, operational and tactical levels and their impact on the architecture and solutions;
  • The provision of secure and resilient services, multiple levels of security, hardware and software certification, cyber protection, data integrity solutions, etc.;
  • The complementarity and synergies, avoiding duplication but bridging existing/upcoming single-domain cloud-based solutions, creating synergies between other already or soon-to-be launched cloud-based important initiatives and enhancing these efforts by enabling an advanced use of services and information across domains;
  • The digital continuity aspect, i.e. a virtual environment enabling collaborative operations services across all domains and levels (strategical, operational and tactical);
  • The need for an open architecture, designed in a modular way in order to accommodate specific requirements from tactical level to the different headquarter levels;
  • The need to analyse and compare possible approaches to share data and services within military organisations, depending on the operation levels;
  • The promotion of European standards regarding interoperability and information sharing, and the compatibility with other existing interface military standards such as NATO Federated Mission Networking (FMN);
  • The synergies with European civilian cloud technologies where applicable;

The synergies with the existing cyber ranges in the EU Member States.

The proposals must cover the following activities as referred in article 10.3 of the EDF Regulation, not excluding possible upstream and downstream activities eligible for development actions if deemed useful to reach the objectives:

  • Studies, such as feasibility studies to explore the feasibility of new or improved technologies, products, processes, services and solutions;
  • The design of a defence product, tangible or intangible component or technology as well as the definition of the technical specifications on which such design has been developed which may include partial tests for risk reduction in an industrial or representative environment;

The proposals must address in particular the following tasks:


  • Detailed Requirements Review (DRR), analysing the main operational requirements from Member States at strategic, operative and tactical levels in terms of multi-domain operations;
  • Definition of the Concept of Employment (CONEMP);
  • Definition of use cases, defining actors, ways of operation, time constraints, expected data to be shared, existing or predicted services to be considered
  • High-level feasibility study, identifying the main architecture options and their constraints (centralised /decentralised, cloud federation principles, multi-level security, multi-tenants, national sovereignty, including the possibilities of further utilising the existing Cyber Ranges in the EU for AI sandboxing, …);
  • Definition of Governance;
  • Definition of a delivery model for the building blocks (who delivers what to whom and who supports);
  • IT Platform preparation – small-scale version of the virtual platform to develop and to experiment the federated services (limited to a selection of services);
  • IT & Core Services Modules development and integration in the virtual or digital platform;
  • Development and integration of the demo version of the catalogue; use of a limited set of services in areas to be defined (e.g. C2 or ISR – for illustration purposes only);
  • Technological demonstration of the federating platform and its key services.


  • Conceptual design;
  • Architecture definition, laying down the overall multi-domain federated cloud architecture, including by partially using the existing infrastructure;
  • System of systems specification, outlining platform related services;
  • Functional description, providing first version of the catalogue and the associated interfaces and APIs;
  • Standards and interoperability assessment;
  • Preliminary Design Tool to elaborate and customize architecture for each Member State.
Check the call here