An Application of ELISA to Perfect Hashing with Deterministic Ordering

This paper describes a practical application of a novel terminal attractor algorithm to the construction of Perfect Hash Functions (PHF) for a prede ned set of keys. The proposed method, which can be used in the ordering phase of a classical Mapping, Deterministic Ordering and Searching (MDOS) approach, is based on a neural network canonical form of efcient gradient descent, for solving linear systems, named ELISA. Numerical experiments clearly show that ELISA is able to determine the optimal solution in many di cult cases, where other alternative algorithms fail or are ine ective.

• Type: Conference
• Year: 1998
• Where: ICONIP
• Authors: Monica Bianchini, Stefano Fanelli, Marco Gori
• Keywords: Efcient Gradient Descent, ICONIP 1998, ICONIP 2007, Perfect Hash Functions, Terminal Attractor Algorithm