Neural networks have become increasingly adopted in society over the last few years. As neural networks consume a lot of energy to train, reducing the energy consumption of these networks is desirable from an environmental perspective. Spiking neural network is a type of neural network inspired by the human brain which is significantly more energy efficient than traditional neural networks. However, there is little research about how the hyper parameters of these networks affect the relationship between accuracy and energy. The aim of this report is therefore to analyse this relationship. To do this, we measure the energy usage of training several different spiking network models. The results of this study shows that the choice of hyper-parameters in a neural network does affect the efficiency of the network. While correlation between any individual factors and energy consumption is inconclusive, this work could be used as a springboard for further research in this area. / Under de senaste åren har neuronnät blivit allt vanligare i samhället. Eftersom neuronnät förbrukar mycket energi för att träna dem är det önskvärt ur miljösynpunkt att minska energiförbrukningen för dessa nätverk. Spikande neuronnät är en typ av neuronnät inspirerade av den mänskliga hjärnan som är betydligt mer energieffektivt än traditionella neuronnät. Det finns dock lite forskning om hur hyperparametrarna i dessa nätverk påverkar sambandet mellan noggrannhet och energi. Syftet med denna rapport är därför att analysera detta samband. För att göra detta mäter vi energiförbrukningen vid träning av flera olika modeller av spikande neuronnät-modeller. Resultaten av denna studie visar att valet av hyperparametrar i ett neuronnät påverkar nätverkets effektivitet. Även om korrelationen mellan enskilda faktorer och energiförbrukning inte är entydig kan detta arbete användas som en startpunkt för ytterligare forskning inom detta område.

An acetonitrile based protein extraction method was developed that demonstrated high efficient and effective removal of high abundant proteins from both human and murine serum. The protein content of the extract was characterised using gel electrophoresis, the Bradford assay and liquid chromatography tandem mass spectrometry (LC-MS/MS) with database searching. Selected reaction monitoring (SRM) analysis was used to quantify the levels of high abundant serum proteins to further validate the extraction methodology. The ACN depletion method, in combination with artificial neural networks (ANNs) data mining software, was applied to a murine growth hormone (GH) gene doping study with the aim of identifying biomarker ions capable of detecting gene doping. The LC-MS and ANNs analysis approach failed to conclusively identify a biomarker to gene doping in the mouse model. However, the application of the same technique to serum from a rhGH administration study in humans, returned models capable of discriminating between rhGH treated placebo states. The ion identified as being the most discriminatory was characterised using mass spectrometry, and was derived from the protein leucine-rich a-2-glycoprotein (LRG). Multiple LRG related tryptic peptides were identified as being up-regulated upon dosing with recombinant human GH (rhGH). A high throughput LC-MS/MS and SRM approach was developed to quantify proteins in human serum. The approach was validated by comparison of LC-MS/MS derived APO A1 concentrations with those obtained using established clinical analyser technologies. The LC-MS/MS methodology was applied to a large cohort of 257 serum samples from two rhGH administration studies performed at Royal Free Hospital . The two administrations included serum samples from 15 individuals who had been dosed daily with rhGH. Serum concentrations of the established rhGH biomarker insulin-like growth factor-I (IGF-I) were quantified by LC-MS/MS and compared well with those determined using two different immunoassay-based methodologies. Serum concentrations of the LRG protein were measured simultaneously with IGF-I and appeared to increase in 14 of the 15 rhGH dosed individuals. Combining the LRG and IGF-I data further increased the separation of rhGH treated and placebo states within each individual, and the application of ANNs analysis showed that the combination of the two proteins increased the discrimination characteristics over using IGF-I alone. The murine equivalent of the LRG protein was identified and SRM transitions for a tryptically derived peptide were developed, along with transitions for monitoring a peptide from the murine IGF-I protein. These transitions were used to quantify the two proteins in the remaining aliquots from a murine GH gene doping experiment, however neither protein appeared to increase in the GH +ve plasmid samples that were analysed.

One of the most meaningful technical innovations in cycling over the past two decades was the development of mobile power meters. With the ability to measure the physical strain under “real world” outdoor conditions, the knowledge of the demand during cycling has improved enormously. Power output has been described as the most direct measure of intensity during cycling and consequently power meters becomes a popular tool to monitor the training and racing of cyclists. However, only limited research data are available on the utilisation of power meters for performance assessment in the field or the analysis of training data. Therefore, the aims of the thesis were to evaluate the ecological validity of a field test, to provide an extensive insight into the longitudinal training strategies of world-class cyclists and to investigate the effects of interval training in the field at difference cadences. The first study aimed to assess the reproducibility of power output during a 4-min (TT4) and a 20-min (TT20) time-trial and the relationship with performance markers obtained during a laboratory graded exercise test (GXT). Ventilatory and lactate thresholds during a GXT were measured in competitive male cyclists (n = 15; VO2max 67 ± 5 mL . min−1 . kg−1; Pmax 440 ± 38 W ). Two 4- min and 20-min time-trials were performed on flat roads. Strong intraclass-correlations for TT4 (r = 0.98; 95 % CL: 0.92-0.99) and TT20 (r = 0.98; 95 % CL: 0.95-0.99) were observed. TT4 showed a bias ± random error of −0.8 ± 23W or −0.2 ± 5.5%. During TT20 the bias ± random error was −1.8 ± 14 W or 0.6 ± 4.4 %. Both time-trials were strongly correlated with performance measures from the GXT (p < 0.001). Significant differences were observed between power output during TT4 and GXT measures (p < 0.001). No significant differences were found between TT20 and power output at the second lactate-turn-point (LTP 2) (p = 0.98) and respiratory compensation point (RCP) (p = 0.97). In conclusion, TT4 and TT20 mean power outputs are reliable predictors of endurance performance. TT20 was in agreement with power output at RCP and LTP 2. Study two aimed to quantify power output (PO) and heart rate (HR) distributions across a whole season in elite cyclists. Power output and heart rate were monitored for 11 months in ten male (age: 29.1 ± 6.7 y; VO2max: 66.5 ± 7.1 mL . min−1 . kg−1) and one female (age: 23.1y; VO2max: 71.5 mL . min−1 . kg−1) cyclist. In total, 1802 data sets were sampled and divided into workout categories according to training goals. The PO at the RCP was used to determine seven intensity zones (Z1-Z7). PO and HR distributions into Z1-Z7 were calculated for all data and workout categories. The ratio of mean PO to RCP (intensity factor, IF) was assessed for each training session and for each interval during the training sessions (IFINT). Variability of PO was calculated as coefficient of variation (CV ). There was no significant difference in the distribution of PO and HR for the total season (p = 0.15), although significant differences between workout categories were observed (p < 0.001). Compared with PO, HR distributions showed a shift from low to high intensities. IF was significantly different between categories (p < 0.001). The IFINT was related to performance (p < 0.01), although the overall IF for the session was not. Also, total training time was related to performance (p < 0.05). The variability in PO was inversely associated with performance (p < 0.01). In conclusion, HR accurately reflects exercise intensity over a total season or low intensity workouts but is limited when applied to high intensity workouts. Better performance by cyclists was characterised by lower variability in PO, greater training volume and the production of higher exercise intensities during intervals. The third study tested the effects of low-cadence (60 rev . min−1) uphill (Int60) or high-cadence (100 rev . min−1) flat (Int100) interval training on PO during 20 min uphill (TTup) and flat (TTflat) time-trials. Eighteen male cyclists (VO2max: 58.6 ± 5.4 mL . min−1 . kg−1) were randomly assigned to Int60, Int100 or a control group (Con). The interval training comprised of two training sessions per week over four weeks, which consisted of 6 bouts of 5 min at the PO at RCP. For the control group, no interval training was conducted. A two-factor ANOVA revealed significant increases on performance measures obtained from GXT (Pmax: 2.8 ± 3.0 %; p < 0.01; PO and VO2 at RCP: 3.6 ± 6.3 % and 4.7 ± 8.2 %, respectively; p < 0.05; and VO2 at ventilatory threshold: 4.9 ± 5.6 %; p < 0.01), with no significant group effects. Significant interactions between group and the uphill and flat time-trials, pre vs. post-training on time-trial PO were observed (p < 0.05). Int60 increased PO during both, TTup (4.4 ± 5.3 %) and TTflat (1.5 ± 4.5 %), whereas the changes were − 1.3 ± 3.6 %; 2.6 ± 6.0 % for Int100 and 4.0 ± 4.6 %; − 3.5 ± 5.4 % for Con, during TTup and TTflat, respectively. PO was significantly higher during TTup than TTflat (4.4 ± 6.0 %; 6.3 ± 5.6 %; pre and post-training, respectively; p < 0.001). These findings suggest that higher forces during the low-cadence intervals are potentially beneficial to improve performance. In contrast to the GXT, the time-trials are ecologically valid to detect specific performance adaptations.

In today global business market, concepts are moving towards inter-national and inter-firm in contrary to older days. This has made companies explore ways to leverage their supply chains, and to evaluate the role of suppliers in their activities. One result has been the increased outsourcing of activities and performance. This has heightened the role of suppliers ever more.Consequently, in order to maintain competitive in the market and to be able to satisfy the knowledgeable end customer, suppliers must be completely aligned with the core company and its objectives and performances. This goal is only gained through strategic benchmarking towards Supplier Development.With a large number of suppliers to deal with, it won’t be feasible to have partnership relationship with each. The strategic approach is to classify them in a manner suitable for the company and only then start from top of the list setting goals for the most prioritized.Before making any attempt in Supplier Relationship Development, the company must start from within setting the right mindset from top down, inside its own ground. Only then Supplier development will be aligned with the objectives of the company and the program will be integrated and successful.This research sought after ways to develop a strategic and systematic method of dealing with suppliers. This guideline was classified as a step by step map of a Supplier Development program. The study was conducted for Volvo Penta Vara Company, and as part of its Supply Chain Development Program. The findings from the company and literature were compared and analyzed and finally guidelines were developed.This report is the last part of the master’s studies in Industrial Engineering-Logistics at University of Borås. It gave us the opportunity to use what we had learnt during the programme. Considering the broadness of the subject we were after, this report meant a lot of work. However, this was the challenge we wanted and an intriguing journey where we learned a lot along the way.

<p>Denna undersökning syftar till att undersöka nationella skillnader i moralutveckling och innehåll i moraliskt resonemang. Studien omfattar 445 svenska och 148 amerikanska ungdomar i åldern 14-19 år. De svenska fpp är hämtade från högstadium och gymnasium i Stockholmsområdet, och de amerikanska från två katolska privatskolor i New York. Dessa tog del av ett moralfrågeformulär av Gibbs (Gibbs & Widaman, 1982; Socio-moral Reflection Measure; SRM) som är en reviderad version av Kohlbergs intervjumetod för bedömning av en individs moralutveckling och kvalitativa, moraliska synsätt i frågor rörande moral. Vid analysen förelåg inga nationella skillnader i moralutvecklingsnivå. Resultaten visade ej heller på skillnader mellan svenska och amerikanska ungdomars sätt att resonera kring moraliska frågor, utom i två fall, som kan tänkas visa på ett typiskt svenskt och ett typiskt amerikanskt sätt att tänka och försvara etiska ställningstaganden. Resultatet diskuteras utifrån (nämnda) teorier inom ämnesområdet.</p> / <p>The aim of this study is to investigate whether there are National differences in moral development and content of moral reasoning. The study includes 445 Swedish and 148 North American youths in the age of 14-19 years. The Swedish subjects of the study come from an upper level of compulsory school in the area around Stockholm, and the North American subjects of the study come from two private Catholic schools in New York. The subjects answered Gibbs’ Moral questionnaire (Gibbs & Widaman, 1982; Socio-moral Reflections Measure; SRM), which is a revised version of Kohlberg’s survey method for estimating individual moral development and qualitative moral viewpoints. The analysis showed no Nation-bound differences concerning moral development level between the youths. Neither could any differences between Swedish and North American young peoples’ way of reasoning on moral issues be traced, except in two cases, which possibly show a typical Swedish and a typical North American way of thinking and defending ethical standpoints. The result is discussed in relation to mentioned theories of the subject area.</p>

Denna undersökning syftar till att undersöka nationella skillnader i moralutveckling och innehåll i moraliskt resonemang. Studien omfattar 445 svenska och 148 amerikanska ungdomar i åldern 14-19 år. De svenska fpp är hämtade från högstadium och gymnasium i Stockholmsområdet, och de amerikanska från två katolska privatskolor i New York. Dessa tog del av ett moralfrågeformulär av Gibbs (Gibbs &amp; Widaman, 1982; Socio-moral Reflection Measure; SRM) som är en reviderad version av Kohlbergs intervjumetod för bedömning av en individs moralutveckling och kvalitativa, moraliska synsätt i frågor rörande moral. Vid analysen förelåg inga nationella skillnader i moralutvecklingsnivå. Resultaten visade ej heller på skillnader mellan svenska och amerikanska ungdomars sätt att resonera kring moraliska frågor, utom i två fall, som kan tänkas visa på ett typiskt svenskt och ett typiskt amerikanskt sätt att tänka och försvara etiska ställningstaganden. Resultatet diskuteras utifrån (nämnda) teorier inom ämnesområdet. / The aim of this study is to investigate whether there are National differences in moral development and content of moral reasoning. The study includes 445 Swedish and 148 North American youths in the age of 14-19 years. The Swedish subjects of the study come from an upper level of compulsory school in the area around Stockholm, and the North American subjects of the study come from two private Catholic schools in New York. The subjects answered Gibbs’ Moral questionnaire (Gibbs &amp; Widaman, 1982; Socio-moral Reflections Measure; SRM), which is a revised version of Kohlberg’s survey method for estimating individual moral development and qualitative moral viewpoints. The analysis showed no Nation-bound differences concerning moral development level between the youths. Neither could any differences between Swedish and North American young peoples’ way of reasoning on moral issues be traced, except in two cases, which possibly show a typical Swedish and a typical North American way of thinking and defending ethical standpoints. The result is discussed in relation to mentioned theories of the subject area.

This thesis deals with the design and implementation of a feedback control scheme for a DC Hybrid Active Power Filter used to filter harmonics from a Switched Reluctance Motor (SRM) Drive load. Power electronic systems are non-linear & dynamic [1,3,5]. Power electronic systems employ switching circuits to maximize their efficiency at the penalty that switching circuits generate electrical noise called ripple current and voltage or conducted electromagnetic interference (EMI). The ripple current drawn by the power electronic systems needs to be attenuated to an acceptable level. Filters attenuate this to an acceptable level. Traditionally filters with passive inductors and capacitors are used. Active filters contain switching elements in addition to passive inductors and capacitors which reduce overall size of passive components used. Two control approaches, full-state state space, and plain proportional feedback, are evaluated for this filter. Circuit models are simulated in SPICE and mathematical models are simulated in Matlab/Simulink for evaluating these control approaches. Proportional feedback control was chosen for implementation and the reason for this is provided in the thesis. The active filter was tested with chosen feedback control and experimental results were compared with simulation results. Inferences and scope for further work are finally presented.

Afin de mieux comprendre les mécanismes du trafic cellulaire, les processus de transport des substrats vacuolaires à travers le tonoplaste, le stockage des métabolites et leur dégradation, une analyse globale et exhaustive du protéome vacuolaire d'Arabidopsis thaliana a été réalisée. La connaissance de la localisation subcellulaire des protéines permet de mieux comprendre la fonction des organelles et la compartimentation du métabolisme des plantes. Mais la description précise du protéome d'un organite nécessite d'identifier clairement les véritables protéines résidantes du compartiment étudié. Une tâche si précise est complexe puisqu'elle nécessite la mise en place d'une préparation d'organites purs et homogènes. Pour y parvenir, un protocole de purification de vacuoles à partir de protoplastes isolés de cellules en culture sur un gradient de densité de Ficoll a été amélioré. La combinaison de plusieurs approches de protéomique a permis d'identifier les protéines présentes dans les fractions vacuolaires soluble et membranaire de façon quantitative et fonctionnelle. Les différentes approches ont ainsi mis en évidence des associations et mécanismes moléculaires complexes qui régissent les différentes activités vacuolaires. Cette protéothèque de référence constitue une base pour étudier la dynamique du protéome vacuolaire en réponse à plusieurs stress incluant les métaux lourds. Plusieurs méthodes sans a priori et ciblée ont été proposé afin d'étudier l'impact du cadmium sur la vacuole, ce compartiment cellulaire clé de la détoxication.

Snow indicates the potential stored water volume that is an important source of water supply, which has been the most valuable and indispensable natural resource throughout the history of the world. Euphrates and Tigris, having the biggest dams of Turkey, are the two largest trans-boundary rivers that originate in Turkey and pass throughout the water deficit nations Syria, Iran, Iraq and Saudi Arabia bringing life as well as water all their way. Snowmelt runoff originating from the mountains of Eastern Turkey accounts for 60 to 70 % of total annual discharge observed in Euphrates and Tigris. For an optimum operation of the dams, maximizing energy production, mitigation of floods and satisfying water rights, hydrological models which can both simulate and forecast the river discharges of Euphrates and Tigris are needed. In this study a hydrological model, snowmelt runoff model (SRM), is used in conjunction with remote sensing and geographic information systems to forecast the river discharges in the headwaters of Euphrates River, Upper Euphrates Basin. NOAA and MODIS satellite images were used to derive the snow covered area (SCA) information required by SRM. Linear reduction methodologies based on accumulated air temperature, with constant or varying gradient, were developed to get the continuous daily SCA values from the discrete daily satellite images. Temperature and precipitation forecasts were gathered from two different numerical weather prediction models, namely European Center for Medium Range Weather Forecasts (ECMWF) and Mesoscale Model Version 5 (MM5) from Turkish State Meteorological Services. These data sets provided t+24 hour forecasts of both temperature and precipitation. Temperature, precipitation and SCA information are fed into SRM. Discharge forecasts obtained from the model outputs are compared with the observed values. The overall performance of the model was seen as promising. Possible reasons of the mismatches between the forecasted and observed values are searched. Experiences gained throughout the study are summarized and recommendations on further forecast studies are mentioned.

L’Imagerie et la Spectroscopie par Résonance Magnétique du proton (IRM et ¹H SRM) sont des méthodes innovantes en toxicologie. Ces outils permettent l’évaluation de l’impact des xénobiotiques in vivo au niveau structural et métabolique respectivement. Dans cette perspective, les travaux de cette thèse ont été réalisés sur deux modèles murins : (i) des injections répétées de gadodiamide chez le rat entrainant une accumulation de gadolinium dans les noyaux profonds cérébelleux (NCP) ; (ii) une exposition à faible dose(20 μg/kg/jour) pendant la gestation et la lactation au bisphénol A (BPA) et à ses dérivés chlorés qui sont des perturbateurs endocriniens.L’étude des effets des injections répétées du gadodiamide a montré un rehaussement du signal des NCP sur des images pondérées T1 à 6 et 11 semaines post-injections. De manière inédite, il a été mis en évidence que le rehaussement du signal des NCP se maintient jusqu’à 1 an post-injections. Il a été également mesuré une augmentation prolongée de la concentration de créatine totale dans les NCP par la méthode de quantification ERETIC digital. Néanmoins, cette variation n’a pas été confirmée par les autres méthodes de quantification conventionnelles (jMRUI et LCModel). Une comparaison des méthodes de quantification en ¹H SRM (ERETIC digital, jMRUI et LCModel) a été réalisée afin de mettre en évidence l’influence de la correction de la ligne de base dans le traitement des spectres ¹H in vivo.Des perturbations précoces des dérivés chlorés du BPA ont été quantifiés chez les jeunes souris mâles et femelles exposées pendant la période périnatale : (i) des modifications de la composition lipidique hépatique, et cela préalablement à tout changement structural ; (ii) des altérations microstructurales et métaboliques au niveau de l’hippocampe. Une méthode de diffusion spectroscopique a été optimisée puisdes expériences in vivo ont été effectuées afin de caractériser ces anomalies de l’intégrité intracellulaire dans le cerveau des souris mâles. / Magnetic Resonance Imaging and Spectroscopy (MRI and ¹H MRS) are new methods in toxicology. These tools allow the assessment of the impact of xenobiotics at structural and metabolic levels respectively. In this perspective, this thesis presents two studies in animal models: (i) repeated administrations of linear gadolinium-based contrast agent (gadodiamide) in rat leading to a deposition of gadolinium in deep cerebellar nuclei (DCN); (ii) gestational and lactational exposure to bisphenol A (BPA) and its chlorinated derivatives at low dose (20 μg/kg/day) which are endocrine disruptors.The study of repeated injections of gadodiamide showed a hypersignal of DCN on T1-weighed images at 6and 11 weeks post-injections. In an unprecedented way, it was demonstrated that the hypersignal of DCN persisted for 1 year post-injections. A prolonged increase of total creatine concentration in DCN was also measured by the quantitation ERETIC digital method. Nevertheless, this variation was not confirmed by other conventional quantitation methods (jMRUI and LCModel). A comparison of quantitation methods in ¹H MRS (ERETIC digital, jMRUI and LCModel) was performed in order to elucidate the influence of baseline correction in the processing of in vivo ¹H spectra.Early disturbances caused by chlorinated derivatives of BPA were quantified in young male and female mice perinatally exposed: (i) modifications of hepatic lipid composition which precedes any structural change; (ii)microstructural and metabolic alterations in the hippocampus. A spectroscopic diffusion method was optimized, leading to in vivo experiments in order to characterize those abnormalities of intracellular integrity in the brain of male mice.