Explainable Artificial Intelligence for Energy Systems
Work description
The increasing deployment of PEDs, characterised by a high penetration of distributed energy resources and the need for active flexibility management, has driven the development of data-driven digital tools for forecasting, analysis, and optimisation of energy systems. In this context, Machine Learning models are widely used, but their complexity raises significant challenges in terms of interpretability and user trust. XAI therefore plays a key role in explaining the forecasts, indicators, and decisions generated by these models, contributing to improved transparency and acceptance of decision-support tools in complex, multi-agent energy environments such as PEDs. The main activities planned include: - Development and adaptation of energy forecasting models (e.g., renewable generation, demand, flexibility) in the context of PEDs; - Application and comparison of advanced XAI techniques to explain forecasts, performance indicators, and operational decisions; - Integration of XAI methods into data pipelines and existing Machine Learning models; - Critical analysis of interpretability, robustness, and limitations of models and explanation approaches; - Development of advanced visualisations and interpretable interfaces for different types of users; - Contribution to the validation of models and XAI approaches using real or simulated data; - Support in defining evaluation metrics for explanation quality (e.g., fidelity, stability, usefulness); - Preparation of technical reports and active contribution to scientific publications.
Minimum profile required
- Basic knowledge of Artificial Intelligence and Machine Learning;- Fundamental understanding of Explainable AI (XAI);- Programming skills in Python, including the use of scientific libraries (e.g., NumPy, Pandas);- Basic knowledge of data analysis and time-series analysis;- Fluency in English (written and spoken);
Preference factors
- Experience in Artificial Intelligence, with a focus on XAI and Machine Learning; - Knowledge and experience in applying XAI techniques (e.g., SHAP, LIME, feature importance methods, or counterfactual explanations); - Experience in developing forecasting models applied to energy systems (e.g., renewable generation, demand, flexibility); - Programming skills in Python and experience with data analysis and AI libraries (e.g., Pandas, Scikit-learn, PyTorch, TensorFlow); - Experience in integrating Machine Learning models into data pipelines or practical applications; - Knowledge of energy systems, particularly distributed energy resources and flexibility management; - Experience in data visualisation and development of interpretable interfaces; - Strong analytical skills and interest in model evaluation and interpretability metrics;
Application Period
Since 07 May 2026 to 21 May 2026
Centre
Power and Energy Systems