Generalizing smoothness constraints from discrete samples
We study how certain smoothness constraints, for example, piecewise continuity, can be generalized from a discrete set of analog-valued data, by modifying the error backpropagation, learning algorithm. Numerical simulations demonstrate that by imposing two heuristic objectives — (1) reducing the number of hidden units, and (2) minimizing the magnitudes of the weights in the network — during the learning process, one obtains a network with a response function that smoothly interpolates between the training data.
© 1990 Massachusetts Institute of Technology. Received 15 August 1989; accepted 20 February 1990. Posted Online March 13, 2008. We would like to thank our colleagues, Mr. Mark Neifeld, Dr. Jeff Yu, Dr. Fai Mok, and Mr. Alan Yamamura for helpful discussions. This work was supported by the JPL Director's Discretionary Fund and in part by DARPA and the Air Force Office of Scientific Research.
Published - JICnc90.pdf