Exploring constructive cascade networks

Constructive algorithms have proved to be powerful methods for training feedforward neural networks. An important property of these algorithms is generalization. A series of empirical studies were performed to examine the effect of regularization on generalization in constructive cascade algorithms. It was found that the combination of early stopping and regularization resulted in better generalization than the use of early stopping alone. A cubic penalty term that greatly penalizes large weights was shown to be beneficial for generalization in cascade networks. An adaptive method of setting the regularization magnitude in constructive algorithms was introduced and shown to produce generalization results similar to those obtained with a fixed, user-optimized regularization setting. This adaptive method also resulted in the construction of smaller networks for more complex problems. The acasper algorithm, which incorporates the insights obtained from the empirical studies, was shown to have good generalization and network construction properties. This algorithm was compared to the cascade correlation algorithm on the Proben 1 and additional regression data sets.     

Design and fabrication of a 1 m model of the 70 mm bore twinaperture superconducting quadrupole for the LHC insertions

For reasons of geometrical acceptance, 70 mm bore twin aperture quadrupoles are required in the LHC insertions. For an operating gradient of 160 T/m at 4.5 K, a design based on a four layer coil wound from two graded 8.2 mm NbTi conductors has been developed. Three 1 m single aperture quadrupoles of this design have been built and successfully tested. Thereafter, the magnets have been disassembled and the coils re-collared using self-supporting collars. In this paper, we describe the design features of the twin aperture quadrupole, and report on the initial collaring tests and procedures for collaring and final assembly of the 1 m magnet     

Simulated annealing and weight decay in adaptive learning: theSARPROP algorithm

A problem with gradient descent algorithms is that they can converge to poorly performing local minima. Global optimization algorithms address this problem, but at the cost of greatly increased training times. This work examines combining gradient descent with the global optimization technique of simulated annealing (SA). Simulated annealing in the form of noise and weight decay is added to resiliant backpropagation (RPROP), a powerful gradient descent algorithm for training feedforward neural networks. The resulting algorithm, SARPROP, is shown through various simulations not only to be able to escape local minima, but is also able to maintain, and often improve the training times of the RPROP algorithm. In addition, SARPROP may be used with a restart training phase which allows a more thorough search of the error surface and provides an automatic annealing schedule.     

Refurbishment of Five Ancient Weirs on the River Avon

This paper reviews the work which was carried out during the Avon Weirs Project - from feasibility through site investigation and detailed design, to construction. The design included measures to (a) ensure the stability of the structures against uplift and piping failure, and (b) provide protection to the bank and bed of the river. In addition, the design satisfied the requirement to provide fish passes at all weirs and canoe ramps at Evesham and Chadbury weirs. Most of the funding was provided by the River Avon Weirs'Trust.