Curvature-driven smoothing: a learning algorithm for feed-forward networks

The performance of feed-forward neural networks in real applications can be often be improved significantly if use is made of a-priori information. For interpolation problems this prior knowledge frequently includes smoothness requirements on the network mapping, and can be imposed by the addition to the error function of suitable regularization terms. The new error function, however, now depends on the derivatives of the network mapping, and so the standard back-propagation algorithm cannot be applied. In this paper, we derive a computationally efficient learning algorithm, for a feed-forward network of arbitrary topology, which can be used to minimize the new error function. Networks having a single hidden layer, for which the learning algorithm simplifies, are treated as a special case.

Regularization and complexity control in feed-forward networks

In this paper we consider four alternative approaches to complexity control in feed-forward networks based respectively on architecture selection, regularization, early stopping, and training with noise. We show that there are close similarities between these approaches and we argue that, for most practical applications, the technique of regularization should be the method of choice.

Topographic mappings and feed-forward neural networks

This thesis is a study of the generation of topographic mappings - dimension reducing transformations of data that preserve some element of geometric structure - with feed-forward neural networks. As an alternative to established methods, a transformational variant of Sammon's method is proposed, where the projection is effected by a radial basis function neural network. This approach is related to the statistical field of multidimensional scaling, and from that the concept of a 'subjective metric' is defined, which permits the exploitation of additional prior knowledge concerning the data in the mapping process. This then enables the generation of more appropriate feature spaces for the purposes of enhanced visualisation or subsequent classification. A comparison with established methods for feature extraction is given for data taken from the 1992 Research Assessment Exercise for higher educational institutions in the United Kingdom. This is a difficult high-dimensional dataset, and illustrates well the benefit of the new topographic technique. A generalisation of the proposed model is considered for implementation of the classical multidimensional scaling (¸mds}) routine. This is related to Oja's principal subspace neural network, whose learning rule is shown to descend the error surface of the proposed ¸mds model. Some of the technical issues concerning the design and training of topographic neural networks are investigated. It is shown that neural network models can be less sensitive to entrapment in the sub-optimal global minima that badly affect the standard Sammon algorithm, and tend to exhibit good generalisation as a result of implicit weight decay in the training process. It is further argued that for ideal structure retention, the network transformation should be perfectly smooth for all inter-data directions in input space. Finally, there is a critique of optimisation techniques for topographic mappings, and a new training algorithm is proposed. A convergence proof is given, and the method is shown to produce lower-error mappings more rapidly than previous algorithms.

Drying of portland cement raw material slurries:predicting the performance of a counter-current spray drier

An extensive review of literature has been carried out concerning the drying of single drops, sprays of droplets and the prediction of spray drier performances. The experimental investigation has been divided into two broad parts mainly: (1) Single Drop Experiments, and (2) Spray Drying and Residence Time Distribution Experiments. The thermal conductivity of slurry cakes from five different sources have been experimentally determined using a modified Lee's Disc Apparatus and the data collected was correlated by the polynominal... Good agreement was observed between the experimental thermal conductivity values and the predicted ones. The fit gave a variance ... for the various samples experimented on. A mathematical model for estimating crust mass transfer coefficient at high drying temperatures was derived.

The drying of small drops of particulate slurries

The literature on the evaporation of drops of pure liquids, drops containing solids and droplet sprays has been critically reviewed. An experimental study was undertaken on the drying of suspended drops of pure water and aqueous sodium sulphate decahydrate with concentrations varying from 5 to 54. 1 wt. %. Individual drops were suspended from a glass filament balance in a 26 mm I.D. vertical wind tunnel, designed and constructed to supply hot de-humidified air, to simulate conditions encountered in commercial spray driers. A novel thin film thermocouple was developed to facilitate the simultaneous measurement of drop weight and core temperature. The heat conduction through the thermocouple was reduced because of its unique design; using essentially a single 50μ diameter nickel wire. For pure water drops, the Nusselt number was found to be a function of the Reynolds, Prandtl and Transfer numbers for a temperature range between 19 to 79°C.                  Nu = 2 + 0.19 (1/B)0.24 Re0.5 Pr0.33 Two distinct periods were observed during the drying of aqueous sodium sulphate decahydrate. The first period was characterised by the evaporation from a free liquid surface, whilst drying in the second period was controlled by the crust resistance. Fracturing of the crust occurred randomly but was more frequent at higher concentrations and temperatures. A model was proposed for the drying of slurry drops, based on a receding evaporation interface. The model was solved numerically for the variation of core temperature, drop weight and crust thickness as a function of time. Experimental results were in excellent agreement with the model predictions although at higher temperatures modifications to the model had to be made to accommodate the unusual behaviour of sodium sulphate slurries, i.e. the formation of hydrates.