Skip to main content
Aller à la page d’accueil de la Commission européenne (s’ouvre dans une nouvelle fenêtre)
français fr
CORDIS - Résultats de la recherche de l’UE
CORDIS

Physics-informed Spatial Artificial Intelligence

Objectif

There is a broad consensus among the artificial intelligence (AI) community that, while there has been great progress in the field of verbal intelligence through large language models, if AI cannot understand the physical underpinnings of the world, this progress will stagnate. For an AI to understand the world, we need to develop spatial artificial intelligence. By spatial AI we mean artificial intelligence capable of understanding and interacting with three-dimensional, physical environments.

The aim of the proposed research is to develop a spatial AI that, far from depreciating centuries of scientific knowledge about the behavior of our world, incorporates it, thus giving rise to a new discipline: physics-informed spatial AI.

Learned simulation, rooted in computational mechanics, will drive the understanding of the physical scene. To this end, PHYSIA will develop an AI capable of building from scratch digital twins of the objects present in the scene from computer vision. The advantageous use of the laws of physics, through inductive, geometric and learning biases, will allow the models generated here to be predictive and highly accurate.

In turn, generative artificial intelligence that understands the physical workings of the world will enable unprecedented ways of creating and designing new products and processes. A new generation of physics-informed generative adversarial networks will enable the creation of a new generative design methodology.

At the same time, once AI understands how the world works, it will be possible to create synthetic worlds that behave like our own (giving rise to the industrial metaverse) or harness new forms of extended reality to convey knowledge to the engineer in ways not yet explored. Simulated reality (visual, haptic) will be able to interact with physical reality, so that we can explore how new designs interact with the real world without the need for a physical prototype.

Champ scientifique (EuroSciVoc)

Mots‑clés

Programme(s)

Thème(s)

Régime de financement

HORIZON-ERC - HORIZON ERC Grants

Voir tous les projets financés dans le cadre de ce programme de financement

Appel à propositions

(s’ouvre dans une nouvelle fenêtre) ERC-2025-ADG

Voir tous les projets financés au titre de cet appel

Institution d’accueil

UNIVERSIDAD DE ZARAGOZA
Contribution nette de l'UE
€ 2 487 852,00
Coût total
Aucune donnée

Bénéficiaires (1)