We present a novel, nonperturbative extension of Hopfield's Lyapunov function for symmetric neural networks to the generic nonsymmetric case. The resulting nonperturbative condition for asymptotic stability is not simple, but it can be shown to be nonempty by examining the perturbative limit.
» Purchase the full text
Dissertations and theses can be purchased in a variety of formats which may include: PDF for web download, softcover, hardcover, or microform. Click the "Order a Copy" button to see the formats available for this item.
Available without purchase:
Preview
My Research