Table 7. Overview of Combinations of Hyperparameters Explored.

number of layers	parameter	valuesa
4	neurons	(8000, 4000, 1000, X), (4000, 2000, 500, X)
	dropout rate	0, 0.3
	regularizer rate	0.000001
	learning rate	0.0001
5	neurons	(9000, 4000, 1000, 100, X), (5000, 2000, 1000, 100, X)
	regularizer rate	0, 0.0000001
	learning rate	0.0001

^a“X” in the number of neurons denotes the number of end points employed for each multi-task model (i.e., number of neurons in the output layer).