Abstract | ||
---|---|---|
In this paper, we propose an approximation algorithm of one-valued function with respect to the given tutorial data including n-vector inputs, and scalar outputs. We consider the approximation function as the mapping from input space, a subset of Rn, to a surface M, n-dimensional manifold, in Rn+1 By representing the surface by connection of segmented surfaces, we can apply the concept of spline interpolation to the approximation algorithm effectively. We confirm the effectiveness of the proposed method by comparing the network output with those of the non-regularized network and the one with the Gaussian regularizer. |
Year | Venue | Keywords |
---|---|---|
1998 | ICONIP'98: THE FIFTH INTERNATIONAL CONFERENCE ON NEURAL INFORMATION PROCESSING JOINTLY WITH JNNS'98: THE 1998 ANNUAL CONFERENCE OF THE JAPANESE NEURAL NETWORK SOCIETY - PROCEEDINGS, VOLS 1-3 | segmented surface, spline approximation, Hebbian learning rule |
Field | DocType | Citations |
Pattern recognition,Computer science,Artificial intelligence,Artificial neural network | Conference | 0 |
PageRank | References | Authors |
0.34 | 1 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Kuniaki Uto | 1 | 32 | 10.40 |
Yukio Kosugi | 2 | 127 | 26.67 |