Neural nets learning as an inverse problem. (Czech) Zbl 1265.68146

The author formulates the learning of a neural net by means of a suitably defined linear operator in the form of an inverse problem. The use of reproducing-kernel Hilbert spaces enables an elegant characterization of the solution to the learning problem given by a sample of empirical data and offers an improvement of its stability.


68T05 Learning and adaptive systems in artificial intelligence
46E22 Hilbert spaces with reproducing kernels (= (proper) functional Hilbert spaces, including de Branges-Rovnyak and other structured spaces)
47B32 Linear operators in reproducing-kernel Hilbert spaces (including de Branges, de Branges-Rovnyak, and other structured spaces)
65J22 Numerical solution to inverse problems in abstract spaces