Sum–product graphical models

Machine Learning - This paper introduces a probabilistic architecture called sum–product graphical model (SPGM). SPGMs represent a class of probability distributions that combines, for the...     

Flame retardance of insolubilized silica inside of wood material

Beech wood boards were studied with a cone calorimeter. As the specific density of these materials is greater, the total amounts of CO and CO₂ produced were bigger than values measured on spruce specimens. Also, the shape of the heat release rate (HRR) curve was different as its second peak was bigger than the one observed on spruce wood. It was shown that impregnation of wood with water solutions of H₃BO₃ resulted in decreased HRR values with increasing acid concentrations. Mass loss and CO and CO₂ production rates also decrease in the same way. Impregnation with water‐soluble NaOH‐silica resulted in improved data, when compared to untreated specimens, but were not better than those obtained with H₃BO₃. Simple mixing of both fire retardants gave no improvement, nor was the two‐step impregnation with drying effective after the first step. When the second step‐treatment with acid was followed without drying after impregnation with NaOH‐silica in the first step, then improvement was evident. This is explained by precipitation of the silica content inside the material when the NaOH was neutralized with H₃BO₃, which could penetrate inside when the material was not dried before the second step. The improvement resulted in lower smoke production and longer times to ignition. © 2005 Wiley Periodicals, Inc. J Appl Polym Sci 97: 1948–1952, 2005     

Aspects of phosphorus and carbohydrate metabolism in amyotrophic lateral sclerosis

The behaviour of glycaemia, insulinaemia, phosphoraemia, somatotropinaemia,free glycerol and triglyceridaemia was studied in six patients with A.L.S. following sugar load (1 g/Kg) in fasting. The results of glycaemia and insulinaemia were in tune with published data which have pointed to reduced sugar tolerance and reduced insulin secretion in patients with A.L.S. In the present experiments, particularly significant were the phosphoraemia responses. The failure of inorganic phosphorus values to fall after glucose loading suggests that the glycidic intolerance of these patients is related above all to a reduction in functioning muscular mass rather than to insufficient insulin secretion. The reduction in nervous tissue may also be of importance in this sense. In fact, not all biohumoral parameters investigated were similar to those of diabetes because the behaviour of somatotropinaemia, free glycerol and plasma triglycerides was normal. The changed behaviour of phosphorus would thus indicate altered glucose uptake at peripheral tissue level.     

Flammability studies of sodium thiosulphate or metabisulphite impregnated wood using cone calorimeter

Spruce wood boards impregnated with Na₂S₂O₃ or Na₂S₂O₅ were studied with the cone calorimeter. The presence of Na₂S₂O₃ lowered the average heat release rate in comparison to untreated material. The total amounts of CO and CO₂ production were reduced by the treatment and also the specific extinction area and mass loss rate decreased. Washing of the salt from the material with water caused partial loss of the properties. Addition of the second impregnation step using acids (HCOOH, H₃BO₃ or H₃PO₄) resulted in the fixation of the sulphur in wood, but gave not the results of single‐step modification for thiosulphate. With Na₂S₂O₅ and without acid, the CO and time‐to‐ignition values were higher and average heat release rate smaller in comparison to unmodified material. Combined one step Na₂S₂O₅/H₃BO₃ treatment lowered the CO, CO₂ and specific extinction area values in comparison to the modification with Na₂S₂O₅, similarly like it was observed for Na₂S₂O₃/H₃BO₃ two‐step‐treatment. According to time‐to‐ignition values, Na₂S₂O₅ alone at 5% addition is a better flame retardant than Na₂S₂O₃ at 7%, but the effect is diminished at 10% amount or presence of acids. The total smoke release curves showed decrease due to modification in both phases of the process. Introduction of water washing as well as the acid treatment further lowered the values. The best results were achieved with 15% Na₂S₂O₃–2% H₃PO₄–H₂O system. According to the total smoke release curve the specimen produced more than five times smaller amount of smoke than untreated material in the first phase of the process. It seams that the concentration of Na₂S₂O₅ is less affecting the properties than the synergistic effect of the Na₂S₂O₃ or Na₂S₂O₅/H₃BO₃ system. Although the level of smoke is low, the presence of elemental sulphur causes smaller times‐to‐ignition than on starting material. Copyright © 2006 John Wiley & Sons, Ltd.