Currently the decoder is assumed to have full knowledge of the encoding system. However, it is imaginable to have an ML decoder with a suboptimal model of the encoding system. This can be integrated into the method perfectly well, and can help to give insights into such 'imperfect' decoders.
Currently the decoder is assumed to have full knowledge of the encoding system. However, it is imaginable to have an ML decoder with a suboptimal model of the encoding system. This can be integrated into the method perfectly well, and can help to give insights into such 'imperfect' decoders.