Transfer learning in demand response: A review of algorithms for data-efficient modelling and control

How good are TSO load and renewable generation forecasts: Learning curves, challenges, and the road ahead

“…Another emerging, yet important, distinction in forecasting methods is that of global and local forecast models [13]. Local forecast models predict the future for a single time series, while global models can predict several time series simultaneously, thereby potentially improving generalization by leveraging cross-learning across time series [14].…”

“…1. Based on [33], a number of algorithmic remedies are possible, including increasing model order, feature transformations (such as smoothing, normalisation, differencing or log transformations) and use of techniques such as meta-and multi-task learning which are increasingly relevant for forecasting contexts as well [14,43].…”

Energy balances, thermal performance, and heat stress: Disentangling occupant behaviour and weather influences in a Dutch net-zero energy neighborhood

“…This directive facilitated the publication of a new standard, ISO 52016-1, which can be used to assess the overall energy performance of buildings, while striking a compromise between computational costs and simulation accuracy [46,47]. Going forward, such tools are foreseen to help guide the design phase choices to decarbonizing the building stock, but they can also aid in its operational optimization [48].…”

On the contribution of pre-trained models to accuracy and utility in modeling distributed energy resources

“…As a result, several solutions have been proposed in recent years to reduce the data requirements or sample complexity of learning algorithms, including via the use of pre-trained models (PTMs), which rely on simulated or historically observed data, thereby eschewing the need for observational data [1]. More recently, this has also seen significant interest in the modeling and optimization of distributed energy resources [9,10]. In this paper, we dive deeper into the topic to consider how (or whether) PTMs can accelerate the datadriven modelling of distributed energy resources, e.g.…”