post thumbnail

AI-PROFICIENT’s Journey Towards Lifelong Self-Learning in AI

In the dynamic world of artificial intelligence and Industry 4.0, lifelong learning has emerged as a cornerstone for innovation and adaptability. Within the Deliverable D3.5’ Future scenario-based decision making’, Izaskun Fernandez (TEK), Kerman Lopez de Calle (TEK), Pedro de la Peña (IBE), David Martin Barrios (IBE), Alexander Vasylchenko (TF), Chrispijn Keena (TF), Vincent Goossens (TF), Régis Benzmuller (CONTI), Christophe Van Loock (INEOS), Katarina Stanković (IMP), Dea Pujić (IMP), Marc Anderson (UL-ethics), Alexandre Voisin (UL), George Triantafyllou (ATC), and Antonios Mpantis (ATC), shed light on the quest for lifelong self-reinforcement learning capabilities.

This Deliverable plays a pivotal role within the AI-PROFICIENT project, specifically in Task 3.5, which is dedicated to future scenario-based decision-making and lifelong self-learning. Task 3.5 operates within the broader context of WP3: Platform AI Analytics and Decision-Making Support, where the focus is on providing advanced AI capabilities that can learn and adapt throughout their operational lifespan.

To appreciate the significance of this Deliverable we must first look back at some past deliverables that set the stage by outlining end-user requirements related to feedback management, presenting various approaches to human feedback in reinforcement learning, and specifying initial feedback management solutions. They also detailed the feedback management needs concerning evolving AI models within different use cases, all while ensuring ethical and user workload considerations.

As the AI-PROFICIENT team explains, this Deliverable is best viewed as the next chapter in this ongoing narrative. It aligns with Task 4.1, which provides a generic workflow and specific feedback mechanisms, and gathers essential feedback mechanisms while providing the Human-Machine Interface (HMI). D3.5 takes us further into the realm of lifelong self-learning.

Read the full Deliverable here.