The field of energy forecasting and infrastructure monitoring is rapidly advancing with the development of new deep learning models and techniques. Researchers are exploring the use of wavelet transforms, graph attention, and neural ordinary differential equations to improve the accuracy and interpretability of energy forecasting models. Additionally, there is a growing interest in using drive-by vibration response signals for infrastructure health monitoring, with novel frameworks such as WaveletInception-BiLSTM networks showing promising results. The use of asynchronous cross-border market data is also being investigated to improve day-ahead electricity price forecasting in European markets. Noteworthy papers in this area include:
- Wavelet-Enhanced Neural ODE and Graph Attention for Interpretable Energy Forecasting, which introduces a neural framework that integrates continuous-time Neural Ordinary Differential Equations and graph attention for energy forecasting.
- IDS-Net: A novel framework for few-shot photovoltaic power prediction, which proposes a novel interpretable dynamic selection network based on feature information fusion.
- WaveletInception Networks for Drive-by Vibration-Based Infrastructure Health Monitoring, which presents a novel deep learning-based framework for infrastructure health monitoring using drive-by vibration response signals.