Natural language analysis by stochastic optimization: a progress report on Project APRIL

Abstract

Parsing techniques based on rules defining grammaticality are difficult to use with authentic natural-language inputs, which are often grammatically messy. Instead, the APRIL system seeks a labelled tree structure which maximizes a numerical measure of conformity to statistical norms derived from a sample of parsed text. No distinction between legal and illegal trees arises: any labelled tree has a value. Because the search space is large and has an irregular geometry, APRIL seeks the best tree using simulated annealing, a stochastic optimization technique. Beginning with an arbitrary tree, many randomly-generated local modifications are considered and adopted or rejected according to their effect on tree-value: acceptance decisions are made probabilistically, subject to a bias against adverse moves which is very weak at the outset but is made to increase as the random walk through the search space continues. This enables the system to converge on the global optimum without getting trapped in local optima. Performance of an early version of the APRIL system on authentic inputs has been yielding analyses with a mean accuracy upwards of 75%, using a schedule which increases processing linearly with sentence-length; modifications currently being implemented should eliminate many of the remaining errors.     

MEAT TEXTURE CHARACTERISATION: COMPARISON OF CHEWING PATTERNS, SENSORY AND MECHANICAL MEASURES

The texture of beef, varied by muscle (semimembranosus and semitendinosus), myofibrillar status (ageing time and cold-shortening) and cooking temperature (60C and 80C), was determined by a sensory profile, chewing patterns (determined from surface electromyograms) and mechanical properties of the cooked meats measured at low and high deformation in compression, shear (cutting) and by dynamic (viscoelasticity) tests. The meat factors were discriminated differently by the different measurements showing that they detected differently the various origins of meat texture. Only the maximum modulus before 70% compression and sensory assessments and EMG gave discrimination among all the meat factors. Two main sensory dimensions were found from the profile: juiciness and tenderness with a third related to the determination of elasticity. Juiciness, which varied most between cooking temperatures, was better related to mechanical parameters at low strain rather than high strain and to initial muscular activity rather than later muscular activity. Conversely, tenderness was better related to high strain mechanical properties and to measurements in the middle or later part of the mastication process suggesting that tenderness assessment requires structural disintegration of meat.     

Adsorption of Hexavalent Chromium on Chitosan Beads: Sorption Isotherms and Kinetics

Abstract

Chitosan is a biopolymer that is usually obtained in a flaked form, nonporous and partially soluble in acidic media. The low porosity of the polymer introduces diffusion constraints which are rate limiting. Modifying the structure of the chitosan is a way to improve the accessibility of the adsorption sites. In this study, the modifications were carried out by dissolving chitosan flakes in an acetic acid solution followed by precipitation in a sodium hydroxide solution to form gel beads. The study deals with the influence of several parameters (metal ion concentration, size of the beads, Chitosan conditioning) on metal ion diffusion. Both sorption isotherms and kinetics were taken into account. The use of this material allowed an adsorption capacity of 76 mg g^?1 to be obtained.     

HIGH-ORDER CONTRASTS FOR SELF-ADAPTIVE SOURCE SEPARATION

This paper is concerned with the problem of separating independent non-Gaussian sources. This is done by adaptively maximizing a contrast function based on fourth-order cumulants of the (mixed) obser$softhyphen;vations. The first cla ss of solutions involves a first stage where the signal vector is adaptively whitened. In order to implement in the second stage the proper separating task, new contrast functions are proposed, especially when all the source kurtosises have the same sign. These contrasts involve only self-cumulants of the outputs. The second class of solutions requires a single separating stage. However, the associated contrasts involve cross-cumulants in addition to self-cumulants. They essentially apply to correlated ve ctors with normalized powers (rather than to white vectors). The resulting adaptive one-stage and two-stage systems achieve satisfactory separation performance independently of the statistics of sources and of the kind of linear mixture.     

Calorimetric Examination of Cut Fresh Pineapple Metabolism

Isothermal microcalorimetric analysis of cut fresh pineapple was used to examine metabolic events occurring in the fruit under different storage conditions. The results define conditions most suited for cut pineapple storage. The baseline heat production for pineapple tissue is quite low (0.2 μW per mg wet weight). During storage, heat producing metabolic reactions develop which cause spoilage of the fruit. The time course of heat production during storage is affected by changes in atmosphere and by chemical treatments. Effects of inhibitor and antibiotic additions indicate that most of the heat production associated with spoilage is related to the growth of microbial flora.     

纳米技术涌动未来——极其微小的工艺，革命性的效应

技术引擎：碳纳米管 碳纳米管的强度远远超过钢铁，并且拥有极好的导热与导电性，正因如此，其未来的应用涉及面板广，从CPU到电池，再到散热片等。