Schematic representation of the neural network architecture used for classification. Values pass from the input to the output along the connected edges; each node represents a linear combination of the inputs and the application of a non-linear activation function. The value at the output represents the binary classification probability between 0 and 1. The number of input neurons varies for the different configurations (4 for Euclid-only, and 8 after adding ground-based photometry; see Sect. 4). For visualisation, the number of neurons in each hidden layer has been divided by 4.