HUMAINE

Call: HORIZON-CL4-2022-HUMAN-02-01 (A human-centred and ethical development of digital and industrial technologies, 2022)
Project number: 101120218
Type of Action: Horizon Research and Innovation Actions
Granting Authority: European Commission
Duration: 1.10.2023 – 30.9.2026 (36 months)

Website: https://humaine-horizon.eu/

Can humans and Artificial Intelligence (AI) collaborate effectively in dynamic, unstructured environments? This question lies at the heart of the HumAIne project, which aims to create a novel operating system (OS) to enhance the interaction between humans and AI in advanced decision-making. By integrating cutting-edge components—Active Learning (AL), Neuro-Symbolic Learning, Swarm Learning (SL), and eXplainable AI (XAI)—the HumAIne OS places humans at the center, fostering systems that outperform standalone AI and human-only solutions. This innovative platform is designed to drive transformative applications across industries, reshaping how humans and AI address complex challenges together.

Industries face increasing challenges in deploying AI systems that can adapt to the uncertainty and complexity of ever-changing, unstructured environments. Current AI models often fall short in integrating human insights, lack interpretability, and require vast data for training, limiting their effectiveness in dynamic situations. These limitations hinder optimal decision-making and collaboration between humans and AI.

HumAIne addresses these issues by developing a new OS, which integrates four key components:

• Active Learning (AL): Facilitates Human-in-the-Loop systems, enabling human involvement during AI uncertainty for better outcomes.
• Neuro-Symbolic Learning: Merges deep learning with semantic rules, allowing complex tasks to be performed with high accuracy and less training data.
• Swarm Learning (SL): Applies collective intelligence inspired by swarm behavior for robust decision-making in dynamic environments.
• eXplainable AI (XAI): Delivers comprehensive explanations of AI predictions, considering global contexts to guide human operators in optimizing AI systems dynamically.

The unified platform processes diverse data types, correlates human inputs through ontologies and knowledge graphs, and promotes semantic interoperability. In addition to the OS, HumAIne offers training programs and builds a community of stakeholders to drive adoption and maximize the platform’s impact.

HumAIne will deliver a new OS that will integrate Active Learning, Neuro-Symbolic Learning, Swarm Learning, and eXplainable AI to revolutionize Human-AI collaboration. This platform will enhance decision-making across sectors such as Smart Manufacturing, Smart Cities, Energy Grids, Finance, and Healthcare, providing efficient, adaptable, and transparent solutions.

Complementary training resources will empower users to fully utilize the system’s potential, while a vibrant community of stakeholders will promote the platform’s adoption and long-term sustainability. These achievements will lead to scalable solutions for unstructured, dynamic environments, significantly improving efficiency, accuracy, and adaptability in real-world applications.

The HumAIne consortium comprises 17 partners from 12 countries, including leading universities, research institutions, and industry players. This diverse collaboration ensures a multidisciplinary approach to advancing Human-AI collaboration. For more information about the consortium and their roles, visit the official HumAIne website.