Currently, as described in the paper, a somewhat tricky process is used to calculate the compound score for a given arc. It would be good to measure what is the gain of this complication. Perhaps a standard MLP could work just as well in the current architecture.
Currently, as described in the paper, a somewhat tricky process is used to calculate the compound score for a given arc. It would be good to measure what is the gain of this complication. Perhaps a standard
MLPcould work just as well in the current architecture.