Key points are not available for this paper at this time.
We present an analytic solution to the problem of on-line gradient-descent learning for two-layer neural networks with an arbitrary number of hidden units in both teacher and student networks.Received 14 October 1994DOI:https://doi.org/10.1103/PhysRevLett.74.4337©1995 American Physical Society
Saad et al. (Mon,) studied this question.