DesCartes: Augmented Hybrid Engineering
Research Motivation
One of the main challenges in a people-centric smart city is to develop a new phase in real-time decision-making algorithms for critical urban systems. To empower decision-making processes in this very large and very uncertain system of systems environment of the smart city, extremely accurate modelling methods are needed, which we call intelligent modelling. To answer the limitations of current AI (data availability, responsibility, human-AI interaction, trust), we will develop a hybrid AI that hybridizes Learning, Knowledge and Reasoning methods, with the general ability to reason on knowledge derived from Physics and high-level concepts as well as from low-level patterns in data. Thus, Descartes will develop the disruptive concept of hybrid AI (HAI) to leap forward, beyond the current black-box procedures used in fully data-driven AI, by integrating meaning and semantics in the following ways:
- First, it embraces “analog” and “digital” worlds, combining physics knowledge with AI-based data-driven models, giving rise to the novel Hybrid Twins concept within physics-aware AI, that will be the main building block of intelligent modelling.
- Second, it combines reasoning (if, then…), using high level concepts (e.g. colors, shape…), with the more traditional machine-learning approach. This “hybrid” paradigm enables reducing the need for big data and other resources, in particular to produce intelligent modelling that empowers decision-making technologies, as well as making trustworthy AI easier to achieve, and more generally leads to a responsible Hybrid AI.
- Last but not least, a key aspect of DesCartes is to bridge Hybrid AI with Humans as observers, players and decision-makers, to realize the people-centric vision of smart cities.
In summary, HAI will help us to support smart city critical infrastructures:
- smartly, with less data (knowledge/physics-AI)
- safely, certified and regulated (trustworthy-AI)
- carefully, in a human-centric way (human-AI)
- responsibly, by empowering people (society-AI)
Within this large program, the Augmented Hybrid Engineering work package brings a technological and application dimension to the concept of HAI. Whether it is a question of diagnosis or control, key elements such as the quality and relevance of the information provided, the speed of decision making, the robustness of the operating methods, etc., are criteria that the engineer can provide. This work package will contribute efficient modelling techniques that can exhibit the best of physics and data, associated with technological constraints (on sensing, modelling, controlling, etc.). The knowledge on physics-based models is at the heart of the methodology developed in this work package, enabling more robust decisions with explanations. We will focus on research related to the design and engineering of hybrid twin based simulation and control architectures for urban critical systems. In particular, we will focus on decision making activities such as predictive diagnostics and robust optimal control, that impose constraints, and guide our hypothesis to define best suited solutions. Smart data based hybrid twinning brought by this work package will be at the heart of innovations brought to complex system-of-systems. The complexity lies in the various physical scales, the nonlinear behaviour, the various levels of interaction between system components and the uncertain environment making efficient forecasting a real challenge. Tools developed in this work package are generic and can be applied on many real-life engineering systems. A summary of the key objectives and deliverables of this work package can be found here.