Processing Images from Multiple IACTs in the TAIGA Experiment with Convolutional Neural Networks