Descripción del proyecto
Pasos hacia un aprendizaje automático fiable
El aprendizaje automático (AA o ML, por sus siglas en inglés) está preparado para ayudar a avanzar en la profunda descarbonización del sector energético. Su capacidad para aprender en entornos complejos y aportar soluciones pone al AA en condiciones de transformar drásticamente los sistemas energéticos. Sin embargo, las nuevas normas de verificación de la Unión Europea exigirán que se pueda demostrar la fiabilidad de todo el AA y el aprendizaje por refuerzo utilizados en aplicaciones críticas para la seguridad. El objetivo del proyecto TRUST-ML, financiado con fondos europeos, es elaborar un marco unificado para evaluar la fiabilidad cuantitativa de los modelos de redes neuronales utilizados habitualmente en los sistemas energéticos. El equipo de TRUST-ML utilizará un novedoso método de optimización convexa para evaluar la fiabilidad del AA en términos de rendimiento, solidez e interpretabilidad. También está diseñado para satisfacer las nuevas necesidades de los sistemas energéticos ya existentes.
Objetivo
Deep decarbonization of the energy sector will require massive penetration of stochastic renewable energy resources and an enormous amount of grid asset coordination; this represents a challenging paradigm for power system operators. With its ability to learn in complex environments and provide predictive solutions on fast timescales, machine learning (ML) is posed to help overcome these challenges and dramatically transform power systems in coming decades. Emerging EU verification standards, however, will require that all ML and Reinforcement Learning (RL) used in safety critical applications be demonstrably trustworthy. In this project, we develop a unified framework, known as Trust-ML, for assessing the quantitative trustworthiness of the neural network models commonly used in power systems. Trust-ML uses a novel, convex optimization approach to assess ML trustworthiness across three key dimensions: performance, robustness, and interpretability. The approach is engineered to be scalable, and by design, it generates exact verification guarantees. Furthermore, Trust-ML is designed to meet the emerging needs of actual power systems. In particular, it can verify the performance of multi-agent RL systems in rigorous ways, and its relaxed counterpart can offer tractable, worst-case performance guarantees in the context of online learning. The resulting verification tools will be published as open-source software packages and shared widely with researchers and industry. This project will advance state-of-the-art methods across several interdisciplinary fields, it will help remove the barriers associated with machine learning deployment in power systems, and its outcomes will help push European power grids into competitive spaces. Coming from MIT with advanced training in power systems, the project PI, Samuel Chevalier, is characteristically well-suited to build Trust-ML, and his team of advisors represents a mixture of experts across power, optimization, and learning.
Ámbito científico
- natural sciencescomputer and information sciencessoftware
- engineering and technologyenvironmental engineeringenergy and fuelsrenewable energy
- natural sciencescomputer and information sciencesartificial intelligencemachine learningreinforcement learning
- engineering and technologyelectrical engineering, electronic engineering, information engineeringelectrical engineeringpower engineeringelectric power transmission
- natural sciencescomputer and information sciencesartificial intelligencecomputational intelligence
Palabras clave
Programa(s)
- HORIZON.1.2 - Marie Skłodowska-Curie Actions (MSCA) Main Programme
Régimen de financiación
MSCA-PF - MSCA-PFCoordinador
2800 Kongens Lyngby
Dinamarca