The primary goal of ACHILLES is to foster the development of efficient, trustworthy, and responsible data-driven AI applications by creating a modular framework that ensures AI models are lighter, clearer, and safer. The framework supports an iterative lifecycle with distinct stages, addressing human ethics, data, models, deployment, and human agency. Specific objectives (SOs) guide the project’s strategy, with defined results and key performance indicators (KPIs) to measure success.
SO1
Develop ethical and legal frameworks for AI, ensuring compliance with data protection regulations and incorporating human values into AI systems.
SO2
Establish a framework for high-quality, diverse, and representative data, with an emphasis on fairness, bias reduction, and data interoperability.
SO3
Focus on privacy protection by creating tools for GDPR compliance and privacy-preserving AI techniques, including federated learning and synthetic data models.
SO4
Optimize AI model training and deployment to reduce energy consumption, time requirements, and computational costs, aligned with environmental sustainability goals.
SO5
Enhance transparency, explainability, and robustness in AI models, ensuring trustworthy continuous monitoring and vulnerability assessment.
SO6
Develop a user-friendly, ML-driven Integrated Development Environment (IDE) to support efficient, compliant, and ethical AI development throughout the lifecycle.
ACHILLES aims to impact the AI ecosystem by improving performance, compliance, and transparency while promoting responsible AI development across various sectors.
