This paper explores the use of paraconsistent analysis for assessing neural networks from an explainable AI perspective. This is an early exploration paper aiming to understand whether paraconsistent analysis can be applied for understanding neural networks and whether it is worth further develop the subject in future research. The answers to these two questions are affirmative. Paraconsistent analysis provides insightful prediction visualisation through a mature formal framework that provides proper support for reasoning. The significant potential envisioned is the that paraconsistent analysis will be used for guiding neural network development projects, despite the performance issues. This paper provides two explorations. The first was a baseline experiment based on MNIST for establishing the link between paraconsistency and neural networks. The second experiment aimed to detect violence in audio files to verify whether the paraconsistent framework scales to industry level problems. The conclusion shown by this early assessment is that further research on this subject is worthful, and may eventually result in a significant contribution to the field.

In this article, we evaluate the performance of combining several model compression techniques. The techniques assessed were dark knowledge distillation, pruning, and quantization. We found that in the scenario in which we developed the experiments, classification of chest x-rays, the combination of these three techniques yielded a new model capable of aggregating the individual advantages of each one. In the experiments we used a combination of deep models with 95.05% accuracy, a value higher than that reported in some related works but lower than the state of the art, whose accuracy is 96.39%. The accuracy of the compressed model in turn was 90.86%, a small loss compared to the gain obtained from the reduction, in bytes, in relation to the size of the original model. The size has been reduced from 841MB to 40KB, which opens up the possibility for using deep models in edge computing applications.