Adaptive user modelling in intelligent telephone and email assistants

With the burgeoning complexity and capabilities of modern information appliances and services, user modelling is becoming an increasingly important research area. Simple user profiles already personalise many software products and consumer goods such as digital TV recorders and mobile phones. A user model should be easy to initialise, and it must adapt in the light of interaction with the user. In many cases, a large amount of training data is needed to generate a user model, and adaptation is equivalent to retraining the system. This paper briefly outlines the user modelling problem and work done at BTexact on an intelligent personal assistant (IPA) which incorporates a user profile. We go on to describe FILUM, a more flexible method of user modelling, and show its application to the telephone assistant and email assistant components of the IPA, with tests to illustrate its usefulness.     

Quality of Crude Fish Oil Extracted from Herring Byproducts of Varying States of Freshness

ABSTRACT: Herring byproducts were stored at 2 and 15 °C for up to 72 h. Over time, significant increases of total volatile bases (TVB), histamine, putrescine, cadaverine, and tyramine were detected. However, only tyramine and TVB levels were temperature-dependent. The level of total polyunsaturated fatty acids (PUFAs) was constant. Longer byproducts storage gave rise to an oil with higher levels of free fatty acids, eicosapentaenoic acid (EPA), docosahexaenoic acid (DHA), and total PUFAs, while fluorescent compounds were lower. A higher storage temperature led to oil higher in α-tocopherol and EPA levels and lower in anisidine value. Surprisingly, the oil with the highest content of PUFAs was not produced from the freshest byproducts, and oil with low oxidation products can be extracted from stored byproducts.     

Selective Transport of Bismuth Ions through Supported Liquid Membrane

A new supported liquid membrane (SLM) system was prepared for the selective transport of bismuth ions from the aqueous feed into the aqueous permeate phase. The support of the SLM was a thin porous polypropylene or polyvinylidene fluoride membrane impregnated with diisooctyldithiophosphinic acid (Cyanex 301) as mobile carrier in 4‐chloroacetophenon as organic solvent. Cyanex 301 acts as a highly selective carrier for the uphill transport of bismuth ions through the SLM. In the presence of HNO₃ as a metal ion acceptor in the strip solution, the transport of bismuth ions into the strip side reached 70 % of the initial feed concentration after 3.5 hours. The selectivity and efficiency of bismuth transport from aqueous solutions containing different mixtures of cations were investigated. In the presence of P₂O₇^2– ions as suitable masking agent in the feed solution, the interfering effects of other cations were completely eliminated. The selective transport of bismuth through SLM is superior to liquid‐liquid extraction or through bulk liquid membranes. This is due to the high efficiency. The SLM reduces the solvent requirements, combines extraction and stripping operations in a single process and allows the use of highly selective extractants. The system may be applied to samples containing very low bismuth concentrations.     

Preparation and characterization of chitosan‐based dispersions

Complexation of chitosan in aqueous solutions by low molecular weight electrolytes is one of the simplest methods for the preparation of aqueous chitosan dispersions. In this work, the influence of storage time, sulfate concentration, method of preparation and surfactant content on some properties of the resultant chitosan dispersions (turbidity, viscosity and zeta potential) was analyzed. Turbidimetry was adequate to monitor the formation of particles, while viscometry was suitable to monitor changes in the dispersing phase. An analysis of the properties of these systems, mainly in terms of particle–particle and macromolecule–macromolecule interactions was carried out. Copyright © 2004 Society of Chemical Industry     

Immobilization and activation of vanadium(III) and titanium(III) single‐site catalysts for ethylene polymerization using MgCl2‐based supports

Significant increases in the activity of vanadium(III) amidinate catalysts for ethylene polymerization have been obtained by immobilization on a MgCl₂‐based support prepared by reaction of AlEt₃ with a MgCl₂/ethanol adduct. Catalyst immobilization and activation on this type of support prevents the rapid decay in activity observed under homogeneous polymerization conditions with unsupported catalysts. Stable polymerization activity is also observed with analogous titanium(III) complexes. Polyethylene with narrow molecular weight distribution and spherical particle morphology is obtained without reactor fouling. Copyright © 2005 Society of Chemical Industry     

High char‐yielding poly[acrylonitrile‐co‐(itaconic acid)‐co‐(methyl acrylate)]: synthesis and properties

Polyacrylonitrile terpolymers of various compositions consisting of acrylonitrile (AN), itaconic acid (IA) and methyl acrylate (MA) were synthesized by solution polymerization in dimethylsulfoxide. Increase in concentration of either IA or MA retarded the overall polymerization rate and the polymer molecular weight. The system consisting of AN + MA and varying IA concentration was more prone to retardation in comparison with the system composed of AN + IA with variable MA concentration. The retardation factors were quantified. Minor quantities of MA boost the reactivity of IA in the terpolymer system. The terpolymer was richer in MA vis‐à‐vis the feed. The thermal characteristics of the terpolymer were examined as a function of its composition. In contrast to the copolymer of AN and IA requiring 1–1.5 mol% IA, the terpolymer required an IA content of approximately 2.5 mol% for optimum thermal stability. The polymer with 90 mol% AN, 2.5 mol% IA and 7.5 mol% MA exhibited reasonably good char‐forming characteristics and thermal stability. The overall crystallinity and crystallite size of the polymers were found to decrease on incorporation of the comonomers. The ‘aromatization index’ of the copolymer increased with the temperature of pyrolysis through re‐organization of the tetrahydropyridine ladder structure. Copyright © 2005 Society of Chemical Industry     

Geobase: a simple geographical information system on a personal computer

Spatially distributed data are often encountered in the biological sciences. Representation and analysis of such data requires specific tools. A simple geographical information system is presented, which allows representation and elementary analysis of geographically coded information. The system handles two kinds of data: maps and facts, where map data describe the basis on which the fact data are located. Maps consist of objects described through a set of coordinates, while for facts a coordinate pair is associated with an unlimited number of data records containing five fields: a date, an element from a list, a two-character code, an integer number and a real number. The input data can be displayed interactively on screen by logically combining selection criteria for each field. The facts corresponding to the selected criteria are either displayed as such, or are clustered and displayed as polygons or pies. A short example showing a possible application of the program is presented and advantages as well as limitations are discussed.     

Alphorn: a remote procedure call environment for fault-tolerant,heterogeneous, distributed systems

Alphorn, a software environment for programming distributed computer systems, is described. Programs running on different computers, possibly of different types and running different operating systems, communicate in a client-server relationship by means of remote procedure calls. This efficient construct structures programs neatly. The paper covers distributed process control, Alphorn's object-based programming style, remote procedure calls, the service interface language, configuration, runtime and debugging, and fault tolerance support     

GELYMAC: a Macintosh application for calculating DNA fragment size from gel electrophoresis migration data

A program called GELYMAC takes data on the distances migrated by DNA fragments in a one-dimensional electrophoretic gel and, using a cubic-spline best-fit of marker fragment distance migrated versus molecular size, calculates the molecular sizes of the fragments. Written in the Rascal (Real-time Pascal) programming language, the program runs on the Macintosh family of microcomputers. Rapid entry of marker and experimental fragment migration data is afforded using a scroll bar system adjacent to a graphic representation of a gel. Output includes tabular listing of the data, graphic cartoons of the gel, and the fragment locations and molecular sizes for individual gel lanes, and the calibration curve used in data computations.