GYAN: A methodology for rule extraction from artificial neural networks

Artificial neural network (ANN) learning methods provide a robust and non-linear approach to approximating the target function for many classification, regression and clustering problems. ANNs have demonstrated good predictive performance in a wide variety of practical problems. However, there are strong arguments as to why ANNs are not sufficient for the general representation of knowledge. The arguments are the poor comprehensibility of the learned ANN, and the inability to represent explanation structures. The overall objective of this thesis is to address these issues by: (1) explanation of the decision process in ANNs in the form of symbolic rules (predicate rules with variables); and (2) provision of explanatory capability by mapping the general conceptual knowledge that is learned by the neural networks into a knowledge base to be used in a rule-based reasoning system. A multi-stage methodology GYAN is developed and evaluated for the task of extracting knowledge from the trained ANNs. The extracted knowledge is represented in the form of restricted first-order logic rules, and subsequently allows user interaction by interfacing with a knowledge based reasoner. The performance of GYAN is demonstrated using a number of real world and artificial data sets. The empirical results demonstrate that: (1) an equivalent symbolic interpretation is derived describing the overall behaviour of the ANN with high accuracy and fidelity, and (2) a concise explanation is given (in terms of rules, facts and predicates activated in a reasoning episode) as to why a particular instance is being classified into a certain category.

Gene transfer to plants by diverse species of bacteria

Agrobacterium is widely considered to be the only bacterial genus capable of transferring genes to plants. When suitably modified, Agrobacterium has become the most effective vector for gene transfer in plant biotechnology1. However, the complexity of the patent landscape2 has created both real and perceived obstacles to the effective use of this technology for agricultural improvements by many public and private organizations worldwide. Here we show that several species of bacteria outside the Agrobacterium genus can be modified to mediate gene transfer to a number of diverse plants. These plant-associated symbiotic bacteria were made competent for gene transfer by acquisition of both a disarmed Ti plasmid and a suitable binary vector. This alternative to Agrobacterium-mediated technology for crop improvement, in addition to affording a versatile ‘open source’ platform for plant biotechnology, may lead to new uses of natural bacteria– plant interactions to achieve plant transformation.

Invention and business performance in the tissue-engineering industry

Tissue engineering is a young and interdisciplinary scientific discipline but it offers exciting opportunities to improve the quality of health care for hundreds of thousands of patients. Lured by its potential, several start-up companies, pharmaceutical corporations, and medical device enterprises alike are investing heavily in this sector. Invention is a key driver of competition in this sector. In this study, we aim to explain the variation in inventive output across the different firms in the sector. Our major premise is that firms that forge alliances will be able to tap into the expertise of their partners and thus improve their chances of inventive output. We further argue that alliances that enable technology acquisition or learning will enhance the inventive output of firms more than other kinds of alliances. We measure the inventive output of a company by the number of patents filed. On the basis of a preliminary analysis of seven companies, we find support for the hypotheses. We also argue that, to achieve commercial success, firms need to manage time to market (through alliances or otherwise), have a global outlook, nurture their financial resources, and attain critical mass through mergers.

The nonconceptual gateway to early word learning

My research investigates why nouns are learned disproportionately more frequently than other kinds of words during early language acquisition (Gentner, 1982; Gleitman, et al., 2004). This question must be considered in the context of cognitive development in general. Infants have two major streams of environmental information to make meaningful: perceptual and linguistic. Perceptual information flows in from the senses and is processed into symbolic representations by the primitive language of thought (Fodor, 1975). These symbolic representations are then linked to linguistic input to enable language comprehension and ultimately production. Yet, how exactly does perceptual information become conceptualized? Although this question is difficult, there has been progress. One way that children might have an easier job is if they have structures that simplify the data. Thus, if particular sorts of perceptual information could be separated from the mass of input, then it would be easier for children to refer to those specific things when learning words (Spelke, 1990; Pylyshyn, 2003). It would be easier still, if linguistic input was segmented in predictable ways (Gentner, 1982; Gleitman, et al., 2004) Unfortunately the frequency of patterns in lexical or grammatical input cannot explain the cross-cultural and cross-linguistic tendency to favor nouns over verbs and predicates. There are three examples of this failure: 1) a wide variety of nouns are uttered less frequently than a smaller number of verbs and yet are learnt far more easily (Gentner, 1982); 2) word order and morphological transparency offer no insight when you contrast the sentence structures and word inflections of different languages (Slobin, 1973) and 3) particular language teaching behaviors (e.g. pointing at objects and repeating names for them) have little impact on children's tendency to prefer concrete nouns in their first fifty words (Newport, et al., 1977). Although the linguistic solution appears problematic, there has been increasing evidence that the early visual system does indeed segment perceptual information in specific ways before the conscious mind begins to intervene (Pylyshyn, 2003). I argue that nouns are easier to learn because their referents directly connect with innate features of the perceptual faculty. This hypothesis stems from work done on visual indexes by Zenon Pylyshyn (2001, 2003). Pylyshyn argues that the early visual system (the architecture of the "vision module") segments perceptual data into pre-conceptual proto-objects called FINSTs. FINSTs typically correspond to physical things such as Spelke objects (Spelke, 1990). Hence, before conceptualization, visual objects are picked out by the perceptual system demonstratively, like a finger pointing indicating ‘this’ or ‘that’. I suggest that this primitive system of demonstration elaborates on Gareth Evan's (1982) theory of nonconceptual content. Nouns are learnt first because their referents attract demonstrative visual indexes. This theory also explains why infants less often name stationary objects such as plate or table, but do name things that attract the focal attention of the early visual system, i.e., small objects that move, such as ‘dog’ or ‘ball’. This view leaves open the question how blind children learn words for visible objects and why children learn category nouns (e.g. 'dog'), rather than proper nouns (e.g. 'Fido') or higher taxonomic distinctions (e.g. 'animal').

Requirements for enhancing trust, security and integrity of GNSS location services

The paper describes a number of requirements for enhancing the trust of location acquisition from Satellite Navigation Systems, particularly for those applications where the location is monitored through a remote GNSS receiver. We discuss how the trust of a location acquisition could be propagated to an application through the use of a proposed tamper-­resistant GNSS receiver which quantifies the trust of a location solution from the signaling used (ie. P(Y) code, Galileo SOL, PRS, CS) and provides a cryptographic proof of this to a remote application. The tamper­-resistance state of the receiver is also included in this cryptographic proof.

Critical analysis of high particle number emissions from accelerating compressed natural gas buses

Compressed natural gas (CNG) engines are thought to be less harmful to the environment than conventional diesel engines, especially in terms of particle emissions. Although, this is true with respect to particulate matter (PM) emissions, results of particle number (PN) emission comparisons have been inconclusive. In this study, results of on-road and dynamometer studies of buses were used to derive several important conclusions. We show that, although PN emissions from CNG buses are significantly lower than from diesel buses at low engine power, they become comparable at high power. For diesel buses, PN emissions are not significantly different between acceleration and operation at steady maximum power. However, the corresponding PN emissions from CNG buses when accelerating are an order of magnitude greater than when operating at steady maximum power. During acceleration under heavy load, PN emissions from CNG buses are an order of magnitude higher than from diesel buses. The particles emitted from CNG buses are too small to contribute to PM10 emissions or contribute to a reduction of visibility, and may consist of semivolatile nanoparticles.

Characterization of particle number concentrations and PM2.5 in a school : influence of outdoor air pollution on indoor air

Background, Aim and Scope The impact of air pollution on school children’s health is currently one of the key foci of international and national agencies. Of particular concern are ultrafine particles which are emitted in large quantities, contain large concentrations of toxins and are deposited deeply in the respiratory tract. Materials and methods In this study, an intensive sampling campaign of indoor and outdoor airborne particulate matter was carried out in a primary school in February 2006 to investigate indoor and outdoor particle number (PN) and mass concentrations (PM2.5), and particle size distribution, and to evaluate the influence of outdoor air pollution on the indoor air. Results For outdoor PN and PM2.5, early morning and late afternoon peaks were observed on weekdays, which are consistent with traffic rush hours, indicating the predominant effect of vehicular emissions. However, the temporal variations of outdoor PM2.5 and PN concentrations occasionally showed extremely high peaks, mainly due to human activities such as cigarette smoking and the operation of mower near the sampling site. The indoor PM2.5 level was mainly affected by the outdoor PM2.5 (r = 0.68, p<0.01), whereas the indoor PN concentration had some association with outdoor PN values (r = 0.66, p<0.01) even though the indoor PN concentration was occasionally influenced by indoor sources, such as cooking, cleaning and floor polishing activities. Correlation analysis indicated that the outdoor PM2.5 was inversely correlated with the indoor to outdoor PM2.5 ratio (I/O ratio) (r = -0.49, p<0.01), while the indoor PN had a weak correlation with the I/O ratio for PN (r = 0.34, p<0.01). Discussion and Conclusions The results showed that occupancy did not cause any major changes to the modal structure of particle number and size distribution, even though the I/O ratio was different for different size classes. The I/O curves had a maximum value for particles with diameters of 100 – 400 nm under both occupied and unoccupied scenarios, whereas no significant difference in I/O ratio for PM2.5 was observed between occupied and unoccupied conditions. Inspection of the size-resolved I/O ratios in the preschool centre and the classroom suggested that the I/O ratio in the preschool centre was the highest for accumulation mode particles at 600 nm after school hours, whereas the average I/O ratios of both nucleation mode and accumulation mode particles in the classroom were much lower than those of Aitken mode particles. Recommendations and Perspectives The findings obtained in this study are useful for epidemiological studies to estimate the total personal exposure of children, and to develop appropriate control strategies for minimizing the adverse health effects on school children.

Exposure to particle number, surface area and PM concentrations in pizzerias

The aim of this work was to quantify exposure to particles emitted by wood-fired ovens in pizzerias. Overall, 15 microenvironments were chosen and analyzed in a 14-month experimental campaign. Particle number concentration and distribution were measured simultaneously using a Condensation Particle Counter (CPC), a Scanning Mobility Particle Sizer (SMPS), an Aerodynamic Particle Sizer (APS). The surface area and mass distributions and concentrations, as well as the estimation of lung deposition surface area and PM1 were evaluated using the SMPS-APS system with dosimetric models, by taking into account the presence of aggregates on the basis of the Idealized Aggregate (IA) theory. The fraction of inhaled particles deposited in the respiratory system and different fractions of particulate matter were also measured by means of a Nanoparticle Surface Area Monitor (NSAM) and a photometer (DustTrak DRX), respectively. In this way, supplementary data were obtained during the monitoring of trends inside the pizzerias. We found that surface area and PM1 particle concentrations in pizzerias can be very high, especially when compared to other critical microenvironments, such as the transport hubs. During pizza cooking under normal ventilation conditions, concentrations were found up to 74, 70 and 23 times higher than background levels for number, surface area and PM1, respectively. A key parameter is the oven shape factor, defined as the ratio between the size of the face opening in respect

Development of a composite line source emission model for traffic interrupted microenvironments and its application in particle number emissions at a bus station

A composite line source emission (CLSE) model was developed to specifically quantify exposure levels and describe the spatial variability of vehicle emissions in traffic interrupted microenvironments. This model took into account the complexity of vehicle movements in the queue, as well as different emission rates relevant to various driving conditions (cruise, decelerate, idle and accelerate), and it utilised multi-representative segments to capture the accurate emission distribution for real vehicle flow. Hence, this model was able to quickly quantify the time spent in each segment within the considered zone, as well as the composition and position of the requisite segments based on the vehicle fleet information, which not only helped to quantify the enhanced emissions at critical locations, but it also helped to define the emission source distribution of the disrupted steady flow for further dispersion modelling. The model then was applied to estimate particle number emissions at a bi-directional bus station used by diesel and compressed natural gas fuelled buses. It was found that the acceleration distance was of critical importance when estimating particle number emission, since the highest emissions occurred in sections where most of the buses were accelerating and no significant increases were observed at locations where they idled. It was also shown that emissions at the front end of the platform were 43 times greater than at the rear of the platform. Although the CLSE model is intended to be applied in traffic management and transport analysis systems for the evaluation of exposure, as well as the simulation of vehicle emissions in traffic interrupted microenvironments, the bus station model can also be used for the input of initial source definitions in future dispersion models.

Primary students' success on the structured number line

Number lines are part of our everyday life (e.g., thermometers, kitchen scales) and are frequently used in primary mathematics as instructional aids, in texts and for assessment purposes on mathematics tests. There are two major types of number lines; structured number lines, which are the focus of this paper, and empty number lines. Structured number lines represent mathematical information by the placement of marks on a horizontal or vertical line which has been marked into proportional segments (Figure 1). Empty number lines are blank lines which students can use for calculations (Figure 2) and are not discussed further here (see van den Heuvel-Panhuizen, 2008, on the role of empty number lines). In this article, we will focus on how students’ knowledge of the structured number line develops and how they become successful users of this mathematical tool.

Characterisation of the tritrophic interactions between tobacco yellow dwarf virus, its vector and host-plants

Tobacco yellow dwarf virus (TbYDV, family Geminiviridae, genus Mastrevirus) is an economically important pathogen causing summer death and yellow dwarf disease in bean (Phaseolus vulgaris L.) and tobacco (Nicotiana tabacum L.), respectively. Prior to the commencement of this project, little was known about the epidemiology of TbYDV, its vector and host-plant range. As a result, disease control strategies have been restricted to regular poorly timed insecticide applications which are largely ineffective, environmentally hazardous and expensive. In an effort to address this problem, this PhD project was carried out in order to better understand the epidemiology of TbYDV, to identify its host-plant and vectors as well as to characterise the population dynamics and feeding physiology of the main insect vector and other possible vectors. The host-plants and possible leafhopper vectors of TbYDV were assessed over three consecutive growing seasons at seven field sites in the Ovens Valley, Northeastern Victoria, in commercial tobacco and bean growing properties. Leafhoppers and plants were collected and tested for the presence of TbYDV by PCR. Using sweep nets, twenty-three leafhopper species were identified at the seven sites with Orosius orientalis the predominant leafhopper. Of the 23 leafhopper species screened for TbYDV, only Orosius orientalis and Anzygina zealandica tested positive. Forty-two different plant species were also identified at the seven sites and tested. Of these, TbYDV was only detected in four dicotyledonous species, Amaranthus retroflexus, Phaseolus vulgaris, Nicotiana tabacum and Raphanus raphanistrum. Using a quadrat survey, the temporal distribution and diversity of vegetation at four of the field sites was monitored in order to assess the presence of, and changes in, potential host-plants for the leafhopper vector(s) and the virus. These surveys showed that plant composition and the climatic conditions at each site were the major influences on vector numbers, virus presence and the subsequent occurrence of tobacco yellow dwarf and bean summer death diseases. Forty-two plant species were identified from all sites and it was found that sites with the lowest incidence of disease had the highest proportion of monocotyledonous plants that are non hosts for both vector and the virus. In contrast, the sites with the highest disease incidence had more host-plant species for both vector and virus, and experienced higher temperatures and less rainfall. It is likely that these climatic conditions forced the leafhopper to move into the irrigated commercial tobacco and bean crop resulting in disease. In an attempt to understand leafhopper species diversity and abundance, in and around the field borders of commercially grown tobacco crops, leafhoppers were collected from four field sites using three different sampling techniques, namely pan trap, sticky trap and sweep net. Over 51000 leafhopper samples were collected, which comprised 57 species from 11 subfamilies and 19 tribes. Twentythree leafhopper species were recorded for the first time in Victoria in addition to several economically important pest species of crops other than tobacco and bean. The highest number and greatest diversity of leafhoppers were collected in yellow pan traps follow by sticky trap and sweep nets. Orosius orientalis was found to be the most abundant leafhopper collected from all sites with greatest numbers of this leafhopper also caught using the yellow pan trap. Using the three sampling methods mentioned above, the seasonal distribution and population dynamics of O. orientalis was studied at four field sites over three successive growing seasons. The population dynamics of the leafhopper was characterised by trimodal peaks of activity, occurring in the spring and summer months. Although O. orientalis was present in large numbers early in the growing season (September-October), TbYDV was only detected in these leafhoppers between late November and the end of January. The peak in the detection of TbYDV in O. orientalis correlated with the observation of disease symptoms in tobacco and bean and was also associated with warmer temperatures and lower rainfall. To understand the feeding requirements of Orosius orientalis and to enable screening of potential control agents, a chemically-defined artificial diet (designated PT-07) and feeding system was developed. This novel diet formulation allowed survival for O. orientalis for up to 46 days including complete development from first instar through to adulthood. The effect of three selected plant derived proteins, cowpea trypsin inhibitor (CpTi), Galanthus nivalis agglutinin (GNA) and wheat germ agglutinin (WGA), on leafhopper survival and development was assessed. Both GNA and WGA were shown to reduce leafhopper survival and development significantly when incorporated at a 0.1% (w/v) concentration. In contrast, CpTi at the same concentration did not exhibit significant antimetabolic properties. Based on these results, GNA and WGA are potentially useful antimetabolic agents for expression in genetically modified crops to improve the management of O. orientalis, TbYDV and the other pathogens it vectors. Finally, an electrical penetration graph (EPG) was used to study the feeding behaviour of O. orientalis to provide insights into TbYDV acquisition and transmission. Waveforms representing different feeding activity were acquired by EPG from adult O. orientalis feeding on two plant species, Phaseolus vulgaris and Nicotiana tabacum and a simple sucrose-based artificial diet. Five waveforms (designated O1-O5) were observed when O. orientalis fed on P. vulgaris, while only four (O1-O4) and three (O1-O3) waveforms were observed during feeding on N. tabacum and the artificial diet, respectively. The mean duration of each waveform and the waveform type differed markedly depending on the food source. This is the first detailed study on the tritrophic interactions between TbYDV, its leafhopper vector, O. orientalis, and host-plants. The results of this research have provided important fundamental information which can be used to develop more effective control strategies not only for O. orientalis, but also for TbYDV and other pathogens vectored by the leafhopper.

Urban, regional, national and global knowledge capital [Guest editorial]

Purpose – The purpose of this paper is to introduce the JKM 2010 annual special issue on knowledge based development (KBD) with reference to the multi-level analysis characteristic of the field. ----- ----- Design/methodology/approach – A description of the knowledge management approach at ESOC (European Space Operations Centre of the European Space Agency) is provided first. At the core of this approach is the breakdown of knowledge in individual technical domains followed by coverage analysis and criticality assessment. Such a framework becomes the reference for best knowledge acquisition, transfer and storage locus identification and subsequent knowledge management practices and guidelines. ----- ----- Findings – KBD provides an integrated framework to account for multidisciplinary analyses and multilevel practices in knowledge capital generation, distribution and utilization. ----- ----- Originality/value – The collection of papers included in the annual special issue on KBD provides a representative, composite view of the research topics and applications concerns in the field. Involving a number of disciplines and levels of analysis, issues ranging from the technological gatekeeper to global knowledge flows show the interdependence of KBD concepts and tools.

The role of the built environment in the creation, cultivation and acquisition of a knowledge-base

This chapter explores the role of the built environment in the creation, cultivation and acquisition of a knowledge base by people populating the urban landscape. It examines McDonald’s restaurants as a way to comprehend the relevance of the physical design in the diffusion of codified and tacit knowledge at an everyday level. Through an examination of space at a localised level, this chapter describes the synergies of space and the significance of this relationship in navigating the global landscape.

An assessment system for evaluation of driving competencies

Automobiles have deeply impacted the way in which we travel but they have also contributed to many deaths and injury due to crashes. A number of reasons for these crashes have been pointed out by researchers. Inexperience has been identified as a contributing factor to road crashes. Driver’s driving abilities also play a vital role in judging the road environment and reacting in-time to avoid any possible collision. Therefore driver’s perceptual and motor skills remain the key factors impacting on road safety. Our failure to understand what is really important for learners, in terms of competent driving, is one of the many challenges for building better training programs. Driver training is one of the interventions aimed at decreasing the number of crashes that involve young drivers. Currently, there is a need to develop comprehensive driver evaluation system that benefits from the advances in Driver Assistance Systems. A multidisciplinary approach is necessary to explain how driving abilities evolves with on-road driving experience. To our knowledge, driver assistance systems have never been comprehensively used in a driver training context to assess the safety aspect of driving. The aim and novelty of this thesis is to develop and evaluate an Intelligent Driver Training System (IDTS) as an automated assessment tool that will help drivers and their trainers to comprehensively view complex driving manoeuvres and potentially provide effective feedback by post processing the data recorded during driving. This system is designed to help driver trainers to accurately evaluate driver performance and has the potential to provide valuable feedback to the drivers. Since driving is dependent on fuzzy inputs from the driver (i.e. approximate distance calculation from the other vehicles, approximate assumption of the other vehicle speed), it is necessary that the evaluation system is based on criteria and rules that handles uncertain and fuzzy characteristics of the driving tasks. Therefore, the proposed IDTS utilizes fuzzy set theory for the assessment of driver performance. The proposed research program focuses on integrating the multi-sensory information acquired from the vehicle, driver and environment to assess driving competencies. After information acquisition, the current research focuses on automated segmentation of the selected manoeuvres from the driving scenario. This leads to the creation of a model that determines a “competency” criterion through the driving performance protocol used by driver trainers (i.e. expert knowledge) to assess drivers. This is achieved by comprehensively evaluating and assessing the data stream acquired from multiple in-vehicle sensors using fuzzy rules and classifying the driving manoeuvres (i.e. overtake, lane change, T-crossing and turn) between low and high competency. The fuzzy rules use parameters such as following distance, gaze depth and scan area, distance with respect to lanes and excessive acceleration or braking during the manoeuvres to assess competency. These rules that identify driving competency were initially designed with the help of expert’s knowledge (i.e. driver trainers). In-order to fine tune these rules and the parameters that define these rules, a driving experiment was conducted to identify the empirical differences between novice and experienced drivers. The results from the driving experiment indicated that significant differences existed between novice and experienced driver, in terms of their gaze pattern and duration, speed, stop time at the T-crossing, lane keeping and the time spent in lanes while performing the selected manoeuvres. These differences were used to refine the fuzzy membership functions and rules that govern the assessments of the driving tasks. Next, this research focused on providing an integrated visual assessment interface to both driver trainers and their trainees. By providing a rich set of interactive graphical interfaces, displaying information about the driving tasks, Intelligent Driver Training System (IDTS) visualisation module has the potential to give empirical feedback to its users. Lastly, the validation of the IDTS system’s assessment was conducted by comparing IDTS objective assessments, for the driving experiment, with the subjective assessments of the driver trainers for particular manoeuvres. Results show that not only IDTS was able to match the subjective assessments made by driver trainers during the driving experiment but also identified some additional driving manoeuvres performed in low competency that were not identified by the driver trainers due to increased mental workload of trainers when assessing multiple variables that constitute driving. The validation of IDTS emphasized the need for an automated assessment tool that can segment the manoeuvres from the driving scenario, further investigate the variables within that manoeuvre to determine the manoeuvre’s competency and provide integrated visualisation regarding the manoeuvre to its users (i.e. trainers and trainees). Through analysis and validation it was shown that IDTS is a useful assistance tool for driver trainers to empirically assess and potentially provide feedback regarding the manoeuvres undertaken by the drivers.

Translating competency statements into desired outcomes : a representational model for conceptualising the competency acquisition process

The need to develop effective and efficient training programs has been recognised by all sectors engaged in training. In responding to the above need, focus has been directed to developing good competency statements and performance indicators to measure the outcomes. Very little has been done to understand how the competency statements get translated into good performance. To conceptualise this translation process, a representational model based on an information processing paradigm is proposed and discussed. It is argued that learners’ prior knowledge and the effectiveness of the instructional material are two variables that have significant bearing on how effectively the competency knowledge is translated into outcomes. To contextualise the model examples from apprentice training are used.