Description du projet
La conscience collaborative fait son apparition chez les agents artificiels
Les organismes vivants doivent être à la fois conscients d’eux-mêmes et conscients des autres organismes pour faire preuve de comportements coopératifs efficaces. En ce qui concerne les applications d’intelligence artificielle, il n’existe aucun cadre approprié qui ne limite pas l’hétérogénéité des agents individuels et qui permette une prise de conscience globale, sans perdre la conscience locale de chaque agent artificiel. Cela entrave le déploiement d’agents artificiels capables de comportements coopératifs complexes. Le projet EMERGE, financé par l’UE, permettra l’émergence d’une conscience collaborative entre collectifs d’êtres artificiels minimaux sans protocoles prédéfinis. Il prendra en charge des systèmes complexes, distribués et flexibles, présentant une collaboration, une autorégulation et une interopérabilité. Le cadre d’EMERGE offrira un aperçu qui favorisera la prise de conscience afin d’optimiser la réalisation des résultats souhaités.
Objectif
Awareness in biological agents has converging definitions when considering local states describing content-related consciousness from an agent-specific perspective. However, it becomes highly debated when it comes to global states. The issue magnifies when considering collectives of artificial agents. Several frameworks exist, all unsatisfactory in the limitations posed to agents’ heterogeneity and disappearance of the local self into an integrated state. Ultimately, existing frameworks are ineffective in explaining, facilitating, and supporting cooperative behaviours in artificial agents. The lack of a compelling theory of global awareness in AI is currently a significant barrier to the effective deployment of artificial agents in the real world. EMERGE tackles this grand challenge by introducing the novel concept of collaborative awareness for collectives of minimal artificial beings. We will investigate how simple agents can develop a representation of their mutual existence, environment, and cooperative behaviour towards the realization of tasks and goals. EMERGE builds on a scenario of artificial beings with no shared language and constrained individual capabilities, which nevertheless leads to high-complexity behaviours at the collective level. Collaborative awareness becomes an emergent process supporting complex, distributed, and loosely coupled systems capable of high degrees of collaboration, self-regulation, and interoperability without pre-defined protocols. EMERGE delivers a philosophical, mathematical, and technological framework that enables us to know how and where to allocate awareness to achieve a goal through the collective optimally. We will demonstrate EMERGE concepts on robotic use cases, with hints of broader applicability of the framework to Internet of Things, pervasive computing, nanotechnologies. We will also investigate the ethical implications of collaborative awareness, focusing on moral responsibility, vulnerabilities, and trust.
Champ scientifique
Programme(s)
- HORIZON.3.1 - The European Innovation Council (EIC) Main Programme
Régime de financement
HORIZON-AG - HORIZON Action Grant Budget-BasedCoordinateur
56126 Pisa
Italie