I read this documentation about EnsembleArray, and it writes:
However, since the neurons represent different dimensions separately, we cannot compute nonlinear interactions between those dimensions.
I’m not sure I understand this claim… Is there a concrete example of what EnsembleArray can’t do while a high dimension ensemble can?
Hi Nrofis,
Because an EnsembleArray can be treated similarly to an Ensemble, this documentation is just intended to make it clear that they are distinct. When an Ensemble is 2 dimensional you can calculate a nonlinear function (e.g., x[0] * x[1]) of the represented values because the neurons are sensitive to both. If you have a 2D EnsembleArray with 2 Ensembles each representing 1 dimension independently there are no neurons that are sensitive to both x[0] and x[1] and so it’s not possible to decode a nonlinear function (such as x[0] * x[1]).
Does that help?