Autoassociative multilayered neural networks were used to validate the seismic shock classification performed by expert seismologists on the basis of somewhat heuristic criteria. It was possible to ...evaluate the degree to which each event in the training data set belonged to each class, independently of the seismologist's suggestions, avoiding errors when heuristic criteria were not adequate. The networks allow recognition of events incorrectly classified, so that they can be reanalyzed by the human expert.< >
This paper deals with one of the main drawbacks of traditional multi-layer perceptron learning strategy, i.e. the slow rate of convergence during the learning phase. In order to mitigate the ...mentioned problem, the learning parameters of the Back-Propagation algorithm, the learning rate and the momentum, considered as characteristic parameters of a filter, are tuned during the network training in accordance with two introduced functions. Some examples showing the suitability of the proposed strategy are reported. The introduced modification of the Back-Propagation algorithm makes faster the multi-layer neural network learning.
This paper deals with the implementation of multivariable adaptive control algorithms by using an available parallel architecture based on Transputers. A parallelization procedure based on ...optimization of the speed-up ratio is proposed. The parallel architecture described in the paper allows the realization of adaptive controllers for systems with fast dynamics. The results obtained are discussed referring to a classical algorithm, pointing out the advantages and the present limits in using such a type of hardware structure.
The AA propose implementing a self-tuning regulator by means of a mixed transputer-network-based archi tecture and DSP devices wi th transversal fil ter structure. These hardware and software allow ...the realization of new adaptive strategies with wich a vast category of systems can be controlled. The AA discuss in detail how to implement the control algorithm. In addition they report applications to demonstrate the great flexibility of the parallel architecture and the regulator performances as well. With this configuration fast-dynamics systems can be controlled at small expense and with reliable hardware support.
Chemotherapy can serve as a stimulus for mobilizing hematopoietic progenitor cells to the peripheral blood for harvest via leukapheresis. Mobilized peripheral blood stem cells (PBSC) support rapid ...hematologic reconstitution after bone marrow aplasia induced by intensive myelosuppressive treatments. Our purpose was to develop effective mobilization regimens allowing collection of large quantities of PBSC. We administered high-dose cyclophosphamide (HDC, 4 gm/m2) or cyclophosphamide (4 gm/m2) plus etoposide (600 mg/m2) (HDCE) in a nonrandomized, sequential fashion to 94 patients with breast cancer, lymphoma, and other malignancies with collection of PBSC via leukapheresis during white blood cell (WBC) recovery from nadir counts. Each apheresis product was analyzed for total nucleated cell number, granulocyte-macrophage colony-forming units (CFU-GM) and CD34+ cells. Twenty-four additional patients with comparable pretreatment characteristics received HDCE plus recombinant human granulocyte colony-stimulating factor (HDCE+G) after chemotherapy through the end of apheresis. Patients receiving HDC were matched for age, sex, and disease but were more heavily pretreated. HDCE was superior to HDC in mean daily CFU-GM and CD34+ yield (p < 0.05), even when groups were adjusted for performance status and amount of prior therapy. HDCE+G led to 3.7 times more CFU-GM and 4.7 times more CD34+ cells than HDCE. Target PBSC yield, defined as > 20 x 10(4) CFU-GM/kg and >4 x 10(8) cells/kg, was achieved by 92% of HDCE+G patients after a median of three aphereses, 56% of HDCE patients after five aphereses, and 16% of HDC patients after six apheresis (p < 0.0001). Prior chemotherapy inversely correlated with the quantity of PBSC harvested regardless of regimen utilized. Our results demonstrate effective chemotherapy regimens for harvesting hematopoietic progenitors in a diverse patient population. HDCE+G produced the highest number of progenitors, suggesting that increasing dose intensity and adding rhG-CSF enhances mobilization. Correlation between cumulative CD34+ and CFU-GM allows real-time flow cytometric analysis of the number of aphereses required to harvest target numbers of PBSC.
A method involving the balanced gains of a linear system and an optimization procedure is developed in order to obtain a reduced-order model from a high-order linear system. The proposed model order ...reduction procedure is compared through some examples with other classical and modern techniques and is found to have good properties, particularly with regard to the frequency error at both low and high frequencies. The algorithm appears to be very fast and numerically stable for high-order systems, as well. The method can be generalized for the approximation of multiple-input-multiple output (MIMO) systems by formalizing the optimization procedure in the time domain instead of using the frequency-domain approach