Towards a Transparent Deep Ensemble Method Based on Multiagent Argumentation

Learning behavioral models by recurrent neural networks with discrete latent representations with application to a flexible industrial conveyor

“…In the Computer Science field, a lot of research effort has been dedicated to the exploration of discrete latent representations within trained networks [16], [17]. In this context, several techniques extract symbolic rules encoded in a feed-forward architecture as the multi-layer perceptron [18], [19]. RNNs instead process symbolic information in a state-ful manner by encapsulating the dynamics of a system through iterated transformations, following embedded transition rules [20].…”