Assessing the contribution of representation to results

In this paper, we make a methodological point concerning the contribution of the representation of the output of a neural network model when using the model to compare to human error performance. We replicate part of Dell, Juliano & Govindjee’s work on modeling speech errors using recurrent networks (Dell et al., 1993). We find that 1) the error patterns reported by Dell et al. do not appear to remain when more networks are used; and 2) some components of the error patterns that are found can be accounted for by simply adding Gaussian noise to the output representation they used. We suggest tha t when modeling error behavior, the technique of adding noise to the output representation of a network should be used as a control to assess to what degree errors may be attributed to t he underlying network.