117 research outputs found
First person – Agathe Chaigne
First Person is a series of interviews with the first authors of a selection of papers published in Journal of Cell Science, helping early-career researchers promote themselves alongside their papers. Agathe Chaigne is first author on ‘ Three-dimensional geometry controls division symmetry in stem cell colonies’, published in JCS. Agathe is a postdoc in the lab of Ewa Paluch at the MRC Laboratory for Molecular Cell Biology (LMCB), University College London, London, UK, investigating the crosstalk between cell division and cell fate transitions during development
Det ulmer under overflaten. Agathe Backer Grøndahl (1847-1907). Genus, sjanger og norskhet
Abstract Ph.D. dissertation at Göteborg University, Sweden, 2008 Author: Camilla Hambro Title: Det ulmer under overflaten. Agathe Backer Grøndahl (1847–1907), genus, sjanger og norskhet English title: What smoulders beneath the surface. Agathe Backer Grøndahl (1847–1907), gender, genre and Norwegianness Language: Norwegian, with an English summary Department: Department of Culture, Aesthetics and Media Series: Skrifter från musikvetenskap, Göteborgs universitet, nr 91, 2008 ISSN 1654-6261 ISBN 978-91-85974-07-8 My dissertation is a problem oriented one. The composer pianist Agathe Backer Grøndahl (1847–1907) has always been a self-evident figure in chapters of Norwegian music history books that cover music, musicians and composers in "Grieg’s shadow". Extensive historical source materials (reception materials connected to Backer Grøndahl’s concerts and compositions; her letters, notebooks, workbooks and sketches) form the basis of the study. The problems studied are what kind of expectations are connected to three intertwined core concepts connected to Backer Grøndahl, namely: "femininity", "genre", and "Norwegianness" (the folk-influenced Norwegian tradition). Musical and material, as well as institutional, sociological and economical, ”Agathe Backer Grøndahl” discourses are examined. The arenas where the three core concepts and discourses unfold, sound, develop, are moulded and understood, are threefold: •Musical culture with which the pianist composer and her compositions interacted from 1866–1903. •Presentations of her in monographs and music history books. •Various attempts to revise this kind of music historiography. The musical analysis starts out from reception materials connected to her performances of her own works and tries to trace the critics’ descriptions in the scores. The works analyzed are: One of her two orchestral compositions, Andante quasi allegretto for piano and orchestra (1869) in sonata form. (I rediscovered this "lost" work in the National Library in Oslo.) The song "To the queen of my heart" (Op. 1/3, 1870), the descriptively titled lyrical piece Wood Nymph’s Dance (1887), the feminist cantata Nytaarsgry (1901) performed at the Scandinavian Women’s Convention 1902 and the singable and (in)famous song "Eventide" Op. 42/7 1899). Performed compositions become performative via the reception materials, hers, the listeners’ and amateur performers "producerly texts" connected to them
Study and Robustness of Filter Bank Multicarrier Modulation : FBMC
Les systèmes multi-porteuses sont utilisés depuis longtemps, et dans de nombreux standards de communication tels qu'ADSL, DVB-T, WiMax et LTE.la modulation OFDM (Orthogonal Frequency Division Multiplexing) est aujourd'hui la technologie dominante. Cependant, l'innovation qu'est l'agrégation de spectres permet aux systèmes de communication de mieux exploiter le spectre radio, aujourd'hui rare, cher et sous-utilisé.Les futures communications 5G devront trouver le moyen d'exploiter ce spectre fragmenté de manière aussi flexible que possible.Le standard PMR (Private Mobile Radio) fait face aux mêmes problèmes vis-à-vis de l'introduction de services large bande dans un spectre déja surchargé.Ces problématiques nécessitent des formes d'onde dont l'occupation spectrale est presque parfaite, afin de limiter au possible les bandes de garde entre différents utilisateurs. L'occupation spectrale d'OFDM n'est tout simplement pas assez bonne pour satisfaire ces impératifs. Parmi les nouvelles formes d'onde prises en considération dans ce cadre, les systèmes FBMC/OQAM (Filter Bank Multicarrier/Offset QAM) sont conçus de manière à fournir une bien meilleure occupation spectrale que les systèmes OFDM, avec un débit optimal, et sans le besoin d'ajouter un préfixe cyclique.Cependant, un des grands inconvénients des systèmes multi-porteuses tels qu'OFDM ou FBMC/OQAM est leur enveloppe non-constante. Les signaux générés présentent de nombreux pics de puissance élevée qui apparaissent quand les sous-porteuses modulées indépendamment puis sommées sont en phase les unes avec les autres.Cela fait que les signaux multi-porteuses sont très sensibles aux non-linéarités des composants électroniques des systèmes de communication, et tout particulièrement à celles de l'amplificateur de puissance (PA) à l'émission. Ces non-linéarités génèrent des distorsions en dedans et en dehors de la bande utile, ce qui crée des remontées spectrales qui viennent dégrader l'occupation spectrale des signaux.Le but de cette thèse est d'évaluer l'impact que le PA peut avoir sur les performances spectrales du signal FBMC/OQAM, et de réduire la sensibilité de la forme d'ondes à ces non-linéarités.Nous basons nos travaux sur des simulations Matlab et des mesures expérimentales, en utilisant le signal OFDM comme référence.Nous commençons par confirmer que le signal FBMC/OQAM a de meilleures performances spectrales que le signal OFDM. Puis nous quantifions l'effet des non-linéarités de l'amplificateur sur les deux signaux.Ensuite, nous proposons une méthode améliorée de contrôle de la dynamique de l'enveloppe du signal, basée sur une technique de précodage, qui a pour but de réduire la sensibilité du signal FBMC/OQAM aux non-linéarités, pour un coût de complexité modeste.Nous étudions les différents paramètres de cette méthode pour en déduire le paramétrage optimal.Enfin, nous présentons des simulations et des mesures de la capacité de cette méthode à réduire les remontées spectrales en dedans et en dehors de la bande utile quand le signal FBMC/OQAM subit les non-linéarités du PA.Multi-carrier systems are well established in many different communication standards such as ADSL, DVB-T, WiMax, and LTE.The dominant technology for broadband communications today is OFDM (Orthogonal Frequency Division Multiplexing).However the introduction of frequency bands aggregation is allowing systems to deal with a spectrum that is scarce, expensive and underutilized.Future 5G communications must find a way to exploit this fragmented spectrum as flexibly as possible.Similar problems are present also for introducing future broadband PMR (Private Mobile Radio) standards in the already crowded PMR spectrum.This requires waveforms with almost perfect spectrum occupation in order to limit the guard frequency band between users.OFDM's spectral occupation is not good enough to fulfill these requirements.Among the considered waveform approaches, FBMC/OQAM (Filter Bank Multicarrier/Offset QAM) systems are designed to provide a much better spectral occupation than OFDM systems, with optimal data rate, and no need for a cyclic prefix.However, a major disadvantage of a multi-carrier system such as OFDM or FBMC/OQAM is the resulting non constant envelope with numerous high power peaks that appear when the independently modulated sub-carriers are added coherently.This results in a high sensitivity to the non-linearities of electronic components, especially to the PA (Power Amplifier).PA non-linearities generate distortions in-band and out-of band, creating spectral regrowth which degrades the spectral occupation of the signals at the transmitter.The aim of this thesis is to evaluate the impact of the PA on the FBMC/OQAM signal's spectral performances, and to reduce the waveform's sensitivity to those non-linearities.Through simulations and experimental measurements, using the OFDM signal as a basis for comparison, we first confirm FBMC/OQAM's better spectral occupation than OFDM, and then quantify the effect of the PA non-linearities on the FBMC/OQAM an OFDM signals.We then propose an improved precoding method for dynamic envelope control, which aims to reduce the FBMC/OQAM signal's sensitivity to PA non-linearities with limited additional complexity. We study the various parameters in order to provide the optimal parameter choice.Finally, we present simulations and measurements of the method's ability to reduce spectral regrowth in and out of band when the FBMC/OQAM signal is subjected to the PA nonlinearities
Variation spatiale du système squelettique alimentaire de la carpe argentée (Hypophthalmichthys molitrix) dans le cadre de son invasion en Amérique du Nord
RÉSUMÉ: La carpe argentée est l'une des quatre espèces de carpes asiatiques envahissantes en Amérique du Nord qui menacent les écosystèmes en raison de leur grande capacité de compétition. Pour mieux comprendre son adaptabilité, les variations de la croissance et de la forme ont été analysées dans deux populations (Rivière Missouri et Illinois) en ciblant la variation du système squelettique alimentaire. L'analyse des différences dans les patrons de variations entre les deux populations a été effectuée à partir de radiographies du crâne des individus et avec la méthode de la morphométrie géométrique en ciblant les zones d'intérêt du système squelettique alimentaire. Les analyses ont été séparées en deux méthodes pour chaque objectif ; la première utilisant l'anatomie crânienne entière et la seconde utilisant les structures du crâne séparées dans le but d'éclairer l'effet des contraintes architecturales sur les résultats de morphométrie géométrique. La comparaison des résultats obtenus entre les deux méthodes utilisées semble suggérer que l'élimination des contraintes architecturales permettraient d'obtenir des résultats significatifs plus clairs. Les résultats mettent en évidence des différences de taille entre les deux populations, et une absence d'allométrie et de dimorphisme sexuel. La rivière de l'Illinois étant plus riche en phytoplancton, les individus de cette population sont plus grands mais les trajectoires ontogéniques des deux populations sont identiques.
Malgré des environnements différents, la comparaison des patrons de covariation n'a pas démontré de différences ce qui suggère un développement robuste et que la plasticité phénotypique développementale n'est pas à l'origine de l'adaptabilité de cette espèce. Par ailleurs, l'étude des patrons d'intégration morphologique a été réalisée à l'aide du coefficient RV qui détermine le degré d'intégration entre deux structures. Ainsi ce coefficient permet d'investiguer si le système squelettique lié à la prise de nourriture est fortement intégré (les structures sont liées fortement entre elles dans le système) ou plutôt modulaire (les structures sont quasi-indépendantes). Les résultats du coefficient RV ont mis en évidence un système modulaire, ainsi les structures liées à la préhension lors de la prise de nourriture représentent un premier module et les structures liées à la mastication représentent un second module. Entre les structures de ces modules, les résultats montrent une forte intégration suggérant une intégration fonctionnelle forte dans les deux populations. Ainsi la carpe argentée semble avoir un développement intégré du système squelettique alimentaire pour les structures ayant des fonctions partagées afin de faire face aux perturbations environnementales. -- Mot(s) clé(s) en français : Espèce envahissante, Comparaison spatiale, Variation phénotypique, Morphologie, Intégration, Développement, Covariation. -- ABSTRACT: Silver carp is one of four invasive Asian carp species in North America that threaten ecosystems owing to their important competitive ability. A comparison between two populations (Illinois and Missouri rivers) was realised to improve knowledge about the Silver carp's success while focusing on the variation of the food processing skeletal system. The objectives were to obtain a better understanding of the growth and morphological variation in different invaded environments. Analysis of shape variation was conducted on X-rays using geometric morphometrics. Analyses were separated in two methods for each objective; (1) using the entire cranial anatomy and (2) using structures separately in order to highlight the effects of architectural constraints. Comparison of results obtained with the two methods seems to show that eliminating architectural constraints allows recognizing significant differences. First, results show size differences between the two populations but no allometry or sexual dimorphism. Illinois River is rich in phytoplankton owing to commercial activity; thus, individuals are bigger, but the ontogenetic trajectories are similar for the two populations. Secondly, covariance patterns were compared between the two populations, demonstrating that there are no significant differences.
Food skeletal system of the two populations covaries similarly even if the environment differs suggesting ontogenetic robustness for this system. Thirdly, integration patterns were studied with the RV coefficients used to calculate the integration degree among structures. Thus, RV coefficients determine if the structures in the alimentary system are strongly cohesive (i.e. highly integrated) or quasi-independent (i.e. modular). Results of the RV coefficients, used to quantify correlation among structures, show no difference between the Illinois and Missouri populations. The structures from the inferior and superior parts of the feeding skeletal system suggest a modular system. Inferior parts are used for chewing and superior parts are used for prehension. In addition, RV coefficients of the structures in the inferior and superior parts suggest that structures in each module are highly integrated. Finally, results suggest a functional integration in this system for the Illinois and Missouri populations. Silver carp seems to have an integrated development of feeding skeletal system for the structures sharing functions to deal with environmental disturbances. -- Mot(s) clé(s) en anglais : Invasive species, Spatial comparison, Phenotypic variation, Morphology, Integration, Development, Covariation
Formes de représentation, impératif d’actualité et enjeux de pouvoir sur les dispositifs numériques : L’exemple de J.K. Rowling et du site pottermore.com
Cet article interroge l’évolution des perceptions de la figure d’auteur par le public du fait de sa mise en scène sur les dispositifs numériques. Ces espaces, privilégiant la participation de chacun et les jeux collectifs, posent la question de l’autorité et des rapports de force se jouant au sein des textes y circulant. Le cas d’étude examiné ici, celui de la figure de J.K. Rowling sur le site internet pottermore.com, nous amène à définir l’architexte numérique comme un archi-auteur, déterminant les caractéristiques de la fonction-auteur, la posture de l’écrivain et son image dans le corps social. Nous interrogeons également la pertinence de la notion de « marque auteur » sur les dispositifs numériques, qui permettent à l’auteur d’être perçu comme une entité inaccessible chargée de l’autorité très spécifique des marques et des entreprises.This article questions the evolution of public perceptions of authors as a result of their presence on digital platforms. These environments, which privilege individual participation and collective games, question the notion of authority and of power relationships at play in the texts circulating there. The case study examined here, that of J.K. Rowling on the website pottermore.com, leads us to define the digital architext as an “archi-author,” determining the characteristics of the author-function, the stance of the author and her image in society. We also interrogate the relevance of considering the author as a “brand” on digital devices that allow the author to be perceived as an inaccessible entity laden with the very specific form of authority that is characteristic of brands and companies
"De kast, Huldren gjør med Kohalen" Om kjønn og andre musikalske identiteter i den norske komponisten Agathe Backer Grøndahls "Huldreslaat"
This article grew out of an interest in the intriguing Norwegian pianist, composer and piano teacher Agathe Backer Grøndahl (1847-1907), and her composition for piano, Huldreslaat. The work received its first performance with the composer herself at the piano in Kristiania 1887. By virtue of being music both in and as gender discourse, the score was in need of further fleshing-out, understanding and analysis. The author starts out from the assumption that gender roles were and are self organizing dynamic systems and that Grøndahl's gender roles worked towards goals that preserved her own interests. Ahistoric gender perspectives do not penetrate the ideological and social complexity in teh cirtics' understanding of femininity, which deviates substantially from the norms of the present. On the other hand close readings of them show us live experiences from the past. It seems like teh delineation of the work's title and Grøndahl's personal appearance were brought to bear on the inherent mednings of her piano playing, and the work in the performance process. Therefore, it is neither possible nor desirable to regard Huldreslaat in total isolation, since gender, Norwegian and international musical identities, are fundamentally intertwined and interdependent in the work, it's context, and reception materials
On the fairness of crowdsourced training data and Machine Learning models for the prediction of subjective properties. The case of sentence toxicity: To be or not to be #$@&%*! toxic? To be or not to be fair?
Training machine learning (ML) models for natural language processing usually requires lots of data that is often acquired through crowdsourcing. In crowdsourcing, crowd workers annotate data samples according to one or more properties, such as the sentiment of a sentence, the violence of a video segment, the aesthetics of an image, ... To ensure quality of the annotations, several workers annotate the same sample, and their annotations are combined into one unique label using aggregation techniques such as majority voting.When the property to be annotated by the workers is subjective, the workers’ annotations for one same sample might differ, but all be valid. The way the annotations are aggregated can have an effect on the fairness of the outputs of the trained model. For example only accounting for the majority vote leads to ignoring the workers’ opinions which differ from the majority and consequently being discriminative towards certain workers. Also, ML models are not always designed to account for individual opinions, for simplicity's or performance's sake. Finally, to the best of our knowledge, no method exists to assess the fairness of a ML algorithm predicting a subjective property. In this thesis we address such limitations by seeking an answer to the following research question: how can targeted crowdsourcing be used to increase the fairness of ML algorithms trained for subjective properties' prediction?We investigate how annotation aggregation via majority voting creates a dataset bias towards the majority opinion, and how this dataset bias in combination with the current limits of ML models lead to an algorithmic bias of the ML models trained with this dataset and unfairness in the model’s outputs. We assume that an ML model able to return each annotation of each user is a fair model. We propose a new evaluation method of the ML models' fairness, and a methodology to highlight and mitigate potential unfairness based on the creation of adapted training datasets and ML models. Although our work is applicable to any kind of label aggregation for any data subject to multiple interpretations, we focus on the effects of the bias introduced by majority voting for the task of predicting sentence toxicity. Our results show that the fairness evaluation method that we create enables to identify unfair algorithms and compare algorithmic fairness, and the final fairness metric is usable in the training process of ML models. The experiments on the models point out that we can mitigate the biases resulting from majority voting and increase the fairness towards the minority opinions. This is provided that the workers’ individual information and each of their annotations are taken into account when training adapted models, rather than only relying on the aggregated annotations, and that the dataset is resampled on criteria according to the favoured aspect of fairness. We also highlight that more work needs to be done to develop crowdsourcing methods to collect high-quality annotations of subjective properties, possibly at low-cost.Computer Science | Web Information System
Parameters Selection for Filter Bank Precoded Filter Bank Multicarrier Systems
International audienc
Price-Level Targeting
In November 2006, the Bank of Canada announced its intention to lead a concerted research program over the next few years on the type of monetary policy framework that would best contribute to the economic well-being of Canadians in the decades ahead. The research will focus on two broad questions: whether economic welfare might be improved by targeting a rate of inflation lower than 2 per cent, and whether economic welfare might be improved by moving from an inflation-targeting (IT) framework to some form of price-level targeting (PLT). This paper focuses on the second question. The author provides an overview of the main conclusions in the literature on the relative merits of replacing IT with PLT, identifies some key outstanding questions, and outlines the Bank's research program. The author concludes that, compared with the conventional wisdom that prevailed a decade ago, recent analysis is more promising for PLT. Nevertheless, the models that have been used so far often ignore some of the key potential benefits, or some of the key potential costs, associated with PLT. More research is needed before one can draw strong conclusions.Monetary policy framework
Learning State Machines to Monitor and Detect Anomalies on a Kubernetes Cluster
These days more companies are shifting towards using cloud environments to provide their services to their client. While it is easy to set up a cloud environment, it is equally important to monitor the system's runtime behaviour and identify anomalous behaviours that occur during its operation. In recent years, the utilisation of Recurrent Neural Networks (RNNs) and Deep Neural Networks (DNNs) to detect anomalies that might occur during runtime has been a trending approach. However, it is unclear how to explain the decisions made by these networks and how these networks should be interpreted to understand the runtime behaviour that they model. On the contrary, state machine models provide an easier manner to interpret and understand the behaviour that they model. In this work, we propose an approach that learns state machine models to model the runtime behaviour of a cloud environment that runs multiple microservice applications. To the best of our knowledge, this is the first work that tries to apply state machine models to microservice architectures. The state machine model is used to detect the different types of attacks that we launch on the cloud environment. From our experiment results, our approach can detect the attacks very well, achieving a balanced accuracy of 99.2% and a F1 score of 0.982. Cyber Securit
- …
