DeepLL: Considering Linear Logic for the Analysis of Deep Learning Experiments