Les tâches de vision artificielle telles que la reconnaissance d’objets demeurent irrésolues à ce jour. Les algorithmes d’apprentissage tels que les Réseaux de Neurones Artificiels (RNA), représentent une approche prometteuse permettant d’apprendre des caractéristiques utiles pour ces tâches. Ce processus d’optimisation est néanmoins difficile. Les réseaux profonds à base de Machine de Boltzmann Restreintes (RBM) ont récemment été proposés afin de guider l’extraction de représentations intermédiaires, grâce à un algorithme d’apprentissage non-supervisé. Ce mémoire présente, par l’entremise de trois articles, des contributions à ce domaine de recherche. Le premier article traite de la RBM convolutionelle. L’usage de champs réceptifs locaux ainsi que le regroupement d’unités cachées en couches partageant les même paramètres, réduit considérablement le nombre de paramètres à apprendre et engendre des détecteurs de caractéristiques locaux et équivariant aux translations. Ceci mène à des modèles ayant une meilleure vraisemblance, comparativement aux RBMs entraînées sur des segments d’images. Le deuxième article est motivé par des découvertes récentes en neurosciences. Il analyse l’impact d’unités quadratiques sur des tâches de classification visuelles, ainsi que celui d’une nouvelle fonction d’activation. Nous observons que les RNAs à base d’unités quadratiques utilisant la fonction softsign, donnent de meilleures performances de généralisation. Le dernière article quand à lui, offre une vision critique des algorithmes populaires d’entraînement de RBMs. Nous montrons que l’algorithme de Divergence Contrastive (CD) et la CD Persistente ne sont pas robustes : tous deux nécessitent une surface d’énergie relativement plate afin que leur chaîne négative puisse mixer. La PCD à "poids rapides" contourne ce problème en perturbant légèrement le modèle, cependant, ceci génère des échantillons bruités. L’usage de chaînes tempérées dans la phase négative est une façon robuste d’adresser ces problèmes et mène à de meilleurs modèles génératifs. / High-level vision tasks such as generic object recognition remain out of reach for modern Artificial Intelligence systems. A promising approach involves learning algorithms, such as the Arficial Neural Network (ANN), which automatically learn to extract useful features for the task at hand. For ANNs, this represents a difficult optimization problem however. Deep Belief Networks have thus been proposed as a way to guide the discovery of intermediate representations, through a greedy unsupervised training of stacked Restricted Boltzmann Machines (RBM). The articles presented here-in represent contributions to this field of research. The first article introduces the convolutional RBM. By mimicking local receptive fields and tying the parameters of hidden units within the same feature map, we considerably reduce the number of parameters to learn and enforce local, shift-equivariant feature detectors. This translates to better likelihood scores, compared to RBMs trained on small image patches. In the second article, recent discoveries in neuroscience motivate an investigation into the impact of higher-order units on visual classification, along with the evaluation of a novel activation function. We show that ANNs with quadratic units using the softsign activation function offer better generalization error across several tasks. Finally, the third article gives a critical look at recently proposed RBM training algorithms. We show that Contrastive Divergence (CD) and Persistent CD are brittle in that they require the energy landscape to be smooth in order for their negative chain to mix well. PCD with fast-weights addresses the issue by performing small model perturbations, but may result in spurious samples. We propose using simulated tempering to draw negative samples. This leads to better generative models and increased robustness to various hyperparameters.

<p>We use common notation &lowast; for distribution (Scshwartz), (M_p) (Beurling) i {M_p} (Roumieu) setting. We introduce and study new (ultra) distribution spaces, the test function spaces&nbsp;<em>D^&lowast;_E</em>&nbsp; and their strong duals <em>D^{<span style="font-size: 10px;">&#39;</span>&lowast;}_E&rsquo;*</em>.These spaces generalize the spaces <em>D^&lowast;_L^q , D&#39;^&lowast;_L^p , B&rsquo;*</em>&nbsp;and their weighted versions. The construction of our new (ultra)distribution &nbsp;spaces is based on the analysis of a suitable translation-invariant Banach space of (ultra)distribution <em>E</em>&nbsp;with continuous translation group, which turns out to be a convolution module over the Beurling algebra&nbsp;<em>L¹_&omega;</em>, where the weight &nbsp;&omega; is related to the translation operators on <em>E</em>.&nbsp;The&nbsp;Banach space&nbsp;<em>E</em>^&rsquo;_&lowast;&nbsp;stands for&nbsp;<em>L¹_&omega;ˇ &lowast; E</em>&rsquo;.&nbsp;We apply our results to the study of the&nbsp;convolution of ultradistributions. The spaces of convolutors&nbsp;<em>O<span style="font-size: 12px;">&rsquo;^&lowast;</span><span style="font-size: 8.33333px;">C</span></em><span style="font-size: 12px;"><em>&nbsp;(</em><strong>R</strong><em>ⁿ)</em>&nbsp;</span>for tempered&nbsp;ultradistributions are analyzed via the duality with respect to the test function<br />spaces<span style="font-size: 12px;">&nbsp;<em>O^&lowast;_C (</em><strong>R</strong><em>ⁿ)</em>,&nbsp;</span>introduced in this thesis. Using the properties of translationinvariant<br />Banach space of ultradistributions <em>E</em> we obtain a full characterization of<br />the general convolution of Roumieu ultradistributions via the space of integrable<br />ultradistributions is obtained. We show: The convolution of two Roumieu ultradistributions&nbsp;<span style="font-size: 12px;"><em>T, S &isin; D&rsquo;^{Mp} (</em><strong>R</strong><em>ⁿ)&nbsp;</em> exists if and only if&nbsp;<em>(</em></span><em>&phi;</em><span style="font-size: 12px;"><em>&nbsp;&lowast; &Scaron;) T &isin; D^&rsquo;{Mp}_L¹(</em><strong>R</strong><em>ⁿ)</em>&nbsp; for every </span><em>&phi;</em><span style="font-size: 12px;"><em>&nbsp;&isin; D ^{Mp} (</em><strong>R</strong><em>ⁿ)</em>.&nbsp;</span>We study boundary values of holomorphic functions defined in tube domains. New edge of the wedge theorems are obtained. The results<br />are then applied to represent<span style="font-size: 12px;">&nbsp;<em>D&rsquo;_E&rsquo;*</em></span><span style="font-size: 12px;">&nbsp;&nbsp;</span>as a quotient space of holomorphic functions.<br />We also give representations of elements of<span style="font-size: 12px;">&nbsp;<em>D&rsquo;_E&rsquo;*</em></span><span style="font-size: 12px;">&nbsp;&nbsp;</span>via the heat kernel method.</p> / <p>Koristimo oznaku &lowast; za distribuciono (Svarcovo), (Mp) (Berlingovo) i&nbsp;{Mp} (Roumieuovo) okruženje. Uvodimo i prouavamo nove (ultra)distribucione&nbsp;prostore, &nbsp;test funkcijske prostore <em>D</em>^&lowast;_E i njihove duale <em>D^&#39;</em>^&lowast;_{<em>E&#39;*</em>}.&nbsp;&nbsp;Ovi prostori uop&scaron;tavaju&nbsp;<br />prostore <em>D</em>^&lowast;_Lq , <em>D</em>^{&#39;&lowast;}_Lp , <em>B^&#39;</em>^&lowast; i njihove težinske verzije. Konstrukcija na&scaron;ih novih&nbsp;<br />(ultra)distribucionih prostora je zasnovana na analizi odgovarajuićh translaciono&nbsp;<br />- invarijantnih Banahovih prostora (ultra)distribucija koje označavamo sa&nbsp;<em>E</em>. Ovi prostori imaju neprekidnu grupu translacija, koja je konvolucioni modul&nbsp;nad &nbsp;Beurlingovom algebrom L¹_&omega;, gde je težina &omega; povezana sa operatorima translacije&nbsp;<br />prostora <em>E</em>. Banahov prostor <em>E^&#39;</em>_{&lowast;&nbsp;}označava prostor <em>L</em>¹_&omega;˅ &lowast; <em>E^&#39;</em>. Koristeći dobijene&nbsp;<br />rezultata proučavamo konvoluciju ultradistribucija. Prostori konvolutora &nbsp;<em>O^&#39;</em>^&lowast;_{<em>C&nbsp;</em>}(<strong>R</strong>ⁿ)&nbsp;temperiranih ultradistribucija, analizirani su pomoću dualnosti&nbsp;<br />test funkcijskih prostora <em>O</em>^&lowast;_<em>C</em> (<strong>R</strong>ⁿ), definisanih u ovoj tezi. Koristeći svojstva&nbsp;<br />translaciono - invarijantnih Banahovih prostora temperiranih ultradistribucija,&nbsp;<br />opet označenih sa <em>E</em>, dobijamo karakterizaciju konvolucije Romuieu-ovih &nbsp;ultradistribucija,&nbsp;preko integrabilnih ultradistribucija. Dokazujemo da: konvolucija&nbsp;<br />dve Roumieu-ove ultradistribucija <em>T</em>, <em>S</em> &isin; <em>D^&#39;</em>^{Mp}&nbsp;(<strong>R</strong>ⁿ) postoji ako i samo ako (&phi; &lowast; <em>S</em>ˇ)<em>T</em> &isin; <em>D^&#39;</em>^{Mp}&nbsp;_L¹ (<strong>R</strong>ⁿ) za svaki &phi; &isin; <em>D</em>^{Mp}(<strong>R</strong>ⁿ). Takođe, proučavamo granične vrednosti holomorfnih funkcija definisanih na tubama. Dokazane su nove teoreme &rdquo;otrog klina&rdquo;. Rezultati se zatim koriste za prezentaciju <em>D^&#39;_E^&#39;</em>_{&lowast;&nbsp;}preko faktor prostora holomorfnih funkcija. Takođe, data je prezentacija elemente <em>D</em>^&#39;_{<em>E^&#39;</em>&lowast;&nbsp;}koristeći heat kernel metode.</p>

O objetivo da presente pesquisa é, decifrando o significado do termo Clavier (-teclado?, em sentido genérico), na obra tecladística de J. S Bach - especialmente nos Prelúdios e Fugas que compõem Das Wohltemperierte Clavier, oder Praeludia und Fugen durch alle Tone und Semitonia - concentrar-se naqueles que apresentam maiores problemas de execução ao cravo e propor uma leitura dessas peças ao órgão de tubos, levando em conta a linguagem instrumental de cada instrumento. / The objective of the present study remains in the meaning of the word Clavier (-keyboard?, in a general sense) in Bach\'s keyboard works - especially the Preludes and Fuges of Das Wolhtemperierte Clavier, oder Praeludia und Fugen durch alle Tone und Semitonia - to focus in those pieces that presents more difficulties to perform on a harpsichord and suggest a pipe organ performance according to each instrumental language.

O objetivo da presente pesquisa é, decifrando o significado do termo Clavier (-teclado?, em sentido genérico), na obra tecladística de J. S Bach - especialmente nos Prelúdios e Fugas que compõem Das Wohltemperierte Clavier, oder Praeludia und Fugen durch alle Tone und Semitonia - concentrar-se naqueles que apresentam maiores problemas de execução ao cravo e propor uma leitura dessas peças ao órgão de tubos, levando em conta a linguagem instrumental de cada instrumento. / The objective of the present study remains in the meaning of the word Clavier (-keyboard?, in a general sense) in Bach\'s keyboard works - especially the Preludes and Fuges of Das Wolhtemperierte Clavier, oder Praeludia und Fugen durch alle Tone und Semitonia - to focus in those pieces that presents more difficulties to perform on a harpsichord and suggest a pipe organ performance according to each instrumental language.

京都大学 / 新制・課程博士 / 博士(理学) / 甲第23003号 / 理博第4680号 / 新制||理||1671(附属図書館) / 京都大学大学院理学研究科物理学・宇宙物理学専攻 / (主査)准教授 福間 將文, 教授 畑 浩之, 教授 田中 貴浩 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DGAM

No description available.

Im Rahmen des 16. Jahreskongresses der GMTH hat ein Workshop stattgefunden, der sich zum Ziel gesetzt hat, hörend die Sensibilität zu wecken für den Klang einer Musik in einer bestimmten Tonart bei einer jeweils historischen Stimmung. Aufgrund der vom 16. bis ins 19. Jahrhundert benutzten ungleichstufigen Temperaturen ›funktionieren‹ nämlich viele Stücke nicht in jeder beliebigen Tonart. Hieraus ergeben sich wichtige Fragen für die Interpretation, für die Komposition und Improvisation, für die Werkanalyse und schließlich für den wissenschaftlichen Diskurs. / During the 16th Annual Conference of the GMTH a workshop was held that aimed to increase the awareness for listening to the sound of a composition in a specific key (within an appropriate historical tuning). Because of the unequal temperaments used from the 16th to the 19th century, many pieces don’t ›work‹ in every key. From this, important questions arise regarding performance, composition, improvisation, analysis, and, finally, scientific discourse.

This thesis comprises two manuscripts prepared for scholarly journals. Chapter 2 comprises an article entitled “Exploring Historic Changes in Musical Communication: Deconstructing Emotional Cues in Preludes by Bach and Chopin.”, which examines emotion perception in historic prelude sets by J.S. Bach and F. Chopin. This work connects psychological research on perceived musical emotion to musicological research describing changes in music structure. Using a technique called commonality analysis to deconstruct cues’ individual and joint roles in predicting participants’ perceived emotions, the chapter clarifies how music’s conveyed emotion can differ in compositions from different eras. Chapter 3 comprises an article entitled “Parsing Musical Patterns in Prelude Sets: Bridging Qualitative and Quantitative Epistemologies in Historical Music Research”. This chapter bridges gaps between qualitative and quantitative research on music history through an analytical approach engaging with both fields. Specifically, cluster analyses of Bach and Chopin’s preludes reveal notable differences in the composers’ expressive toolkits, consistent with work from historical and empirical music research. Through a novel analytical framework, the chapter illustrates a method for detecting groups of pieces demarcated by salient musical differences, assessing cues’ importance within these groups, and determining the most influential cue values for each group. Together, these articles provide new insight into the subtle sonic relationships influencing musical meaning and emotion perception. / Thesis / Master of Science (MSc) / Music’s capacity to express emotion has received considerable attention in psychological and musicological research. Whereas efforts from psychology clarify the musical cues for emotion through perceptual experiments, efforts from musicology track changes in compositional practice over time—finding changing relationships between music’s cues for emotion in historically diverse compositions. To date, the implications of these changing musical relationships for emotion perception remain unclear. This thesis analyzes musical scores and listeners’ emotion ratings to gain insight into music’s structural changes throughout history and their implications for perceived emotion. By applying statistical techniques to (i) detect musical patterns in prelude sets by J.S. Bach and F. Chopin and (ii) clarify how cue relationships influence emotion perception, this thesis sheds light on the relationship between music’s historic context and its emotional meaning.

No description available.

The processes leading to the retention of small quantities of austenite following the bainite and martensite phase transformations have been examined, together with the influence of retained austenite on the properties of low alloy steels. It was found that the upper and lower bainite transformations are separate reactions, although both involve a displacive transformation mode, Growth seems to occur by the repeated nucleation of martensitic sub-units, and this leads to an apparently slow growth rate, The partitioning of carbon from bainitic ferrite into residual austenite was thermodynamically proven to occur subsequent to transformation, and was shown to be directly responsible for the 'incomplete reaction phenomenon'. The nature of sympathetic nucleation and of the limited size of bainitic sub-units was rationalised in terms of the relatively low driving force available for bainite transformations. It was shown that the retention, stability and morphology of austenite could be directly derived from the basic transformation mechanism. Under certain circumstances, the bainitic retained austenite conferred exceptional strength/toughness properties to silicon steels; these were· shown to be superior to the properties associated with tempered martensite microstructures, Using thermodynamics, a model was established which could predict the toughness behaviour of silicon steel bainites simply from a knowledge of the composition. The tempered martensite embrittlement phenomenon was not found to be directly linked to the decomposition of retained austenite films, but to the coarsening of inter- or intra-lath carbides. In dislocated martensites, it was found that the distribution and quantity of retained austenite could be rationalised in terms of the degree of accommodation between adjacent martensite variants. The incipient twins generally observed in lath martensites were shown to be accommodation defects such that the extent of twinning was the greatest when adjacent martensite units had twin-related lattices. The thermodynamics of dislocated martensites have been briefly examined, The inhomogeneous deformation behaviour of dual-phase steels has been analysed in terms of available models.