I would encode DFAs (Deterministic Finite State automata) as output (or input) of a neural network for a supervised learning; it is well-known  that efficacy of neural network training strongly depends on adopted encoding.
How can I encode DFAs for a neural network? Are there any literature works?
I've already found some algorithms being able to extract a DFA from a recurrent neural network, but nothing about DFAs either as input or output of ANN.
: Neural network encoding approach comparison: an empirical study; Investigating the Effect of Data Representation On Neural Network and Regression; Data Mining and Neural Networks: The Impact of Data Representation.