1,720,972 research outputs found
Tessellation-Filtering ReLU Neural Networks
No full textWe identify tessellation-filtering ReLU neural networks that, when composed with another ReLU
network, keep its non-redundant tessellation unchanged or reduce it. The additional network complexity modifies the shape of the decision surface without increasing the number of linear regions. We provide a mathematical understanding of the related additional expressiveness by means of a novel measure of shape complexity by counting deviations from convexity which results in a Boolean algebraic characterization of this special class. A local representation theorem gives rise to novel approaches for pruning and decision surface analysis
Rethinking data augmentation for adversarial robustness
Full text linkRecent work has proposed novel data augmentation methods to improve the adversarial robustness of deep neural networks. In this paper, we re-evaluate such methods through the lens of different metrics that characterize the augmented manifold, finding contradictory evidence. Our extensive empirical analysis involving 5 data augmentation methods, all tested with an increasing probability of augmentation, shows that: (i) novel data augmentation methods proposed to improve adversarial robustness only improve it when combined with classical augmentations (like image flipping and rotation), and even worsen adversarial robustness if used in isolation; and (ii) adversarial robustness is significantly affected by the augmentation probability, conversely to what is claimed in recent work. We conclude by discussing how to rethink the development and evaluation of novel data augmentation methods for adversarial robustness. Our open-source code is available at https://github.com/eghbalz/rethink_da_for_a
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Variations on the Author
Full text link“Variations on the Author” discusses two of Eduardo Coutinho’s recent films (Um Dia na Vida, from 2010, and Últimas Conversas, posthumously released in 2015) and their contribution to the general question of documentary authorship. The director’s filmography is characterized by a consistent yet self-effacing form of authorial self-inscription: Coutinho often features as an interviewer that rather than express opinions propels discourses; an interviewer that is good at listening. This mode of self-inscription characterizes him as an author who is not expressive but who is nonetheless markedly present on the screen. In Um Dia na Vida, however, Coutinho is completely absent form the image, while Últimas Conversas, on the contrary, includes a confessional prologue that moves the director from the margins to the center of his films. This article examines the ways in which these works stand out in the filmography of a director who offers new insights into the notion of cinematic authorship
Appropriate Similarity Measures for Author Cocitation Analysis
Full text linkWe provide a number of new insights into the methodological discussion about author cocitation analysis. We first argue that the use of the Pearson correlation for measuring the similarity between authors’ cocitation profiles is not very satisfactory. We then discuss what kind of similarity measures may be used as an alternative to the Pearson correlation. We consider three similarity measures in particular. One is the well-known cosine. The other two similarity measures have not been used before in the bibliometric literature. Finally, we show by means of an example that our findings have a high practical relevance.information science;Pearson correlation;cosine;similarity measure;author cocitation analysis
Binary Losses for Density Ratio Estimation
Full text linkEstimating the ratio of two probability densities from finitely many
observations of the densities, is a central problem in machine learning and
statistics. A large class of methods constructs estimators from binary
classifiers which distinguish observations from the two densities. However, the
error of these constructions depends on the choice of the binary loss function,
raising the question of which loss function to choose based on desired error
properties. In this work, we start from prescribed error measures in a class of
Bregman divergences and characterize all loss functions that lead to density
ratio estimators with a small error. Our characterization provides a simple
recipe for constructing loss functions with certain properties, such as loss
functions that prioritize an accurate estimation of large values. This
contrasts with classical loss functions, such as the logistic loss or boosting
loss, which prioritize accurate estimation of small values. We provide
numerical illustrations with kernel methods and test their performance in
applications of parameter selection for deep domain adaptation
Dispelling the Myths Behind First-author Citation Counts
Full text linkWe conducted a full-scale evaluative citation analysis study of scholars in the XML research field to explore just how different from each other author rankings resulting from different citation counting methods actually are, and to demonstrate the capability of emerging data and tools on the Web in supporting more realistic citation counting methods. Our results contest some common arguments for the continued
use of first-author citation counts in the evaluation of scholars, such as high correlations between author rankings by first-author citation counts and other citation
counting methods, and high costs of using more realistic citation counting methods that are not well-supported by the ISI databases. It is argued that increasingly available digital full text research papers make it possible for citation analysis studies to go beyond what the ISI databases have directly supported and to employ more
sophisticated methods
Domain Adaptation basierend auf Momenten: Lernschranken und Algorithmen
No full textDiese Dissertation trägt zu den mathematischen Grundlagen des Bereichs "Domain Adaptation" bei, welcher einen aufstrebenden Teilbereich des Maschinellen Lernens bildet. Im Gegensatz zum klassischen Statistischen Lernen berücksichtigt das Framework Domain Adaptation auch Abweichungen zwischen den Wahrscheinlichkeitsverteilungen der Trainings- und Anwendungsumgebung. Domain Adaptation kann damit in breiteren Bereichen eingesetzt werden, da Stichproben zukünftiger Daten oft einer anderen Wahrscheinlichkeitsverteilung folgen als Stichproben der Trainingsdaten. Ein wichtiger Punkt bei Domain Adaptation ist die Allgemeinheit der Annahmen über die Ähnlichkeit der Wahrscheinlichkeitsverteilungen. Aus diesem Grund studieren wir in dieser Dissertation Probleme von Domain Adaptation unter so schwachen Annahmen wie sie mit endlich vielen Momenten modelliert werden können.
Durch die Untersuchung der Generalisierungsfähigkeit von unterscheidenden Modellen, welche unter diesen verallgemeinerten Annahmen gelernt wurden, entwerfen wir im ersten Teil dieser Arbeit ein neues Framework, um obere Schranken für das Missklassifikationsrisiko zu finden. Diese neu beschriebenen oberen Schranken basieren auf endlich vielen Momenten und zusätzlichen Glattheitseigenschaften. Unsere Resultate zeigen, dass ein kleines Missklassifikationsrisiko von unterscheidenden Modellen erwartet werden kann, wenn a) das Missklassifikationsrisiko bezüglich der Trainingsstichprobe klein ist, b) die Stichprobengröße groß genug ist und c) die Wahrscheinlichkeitsverteilungen der Stichproben eine zusätzliche Entropieeigenschaft erfüllen.
Im zweiten Teil setzen wir unser Framework zur Entwicklung neuer Lernalgorithmen ein. Unter Anderem stellen wir eine neue, auf Momenten basierende Distanz für die Regularisierung von Neuronalen Netzen vor. Die von uns vorgestellten Methoden zielen darauf ab, neue Datenrepräsentationen zu finden, welche die im ersten Teil vorgestellten, schwachen Annahmen an die Ähnlichkeit von Wahrscheinlichkeitsverteilungen erfüllen. In diesem Kontext beweisen wir verschiedene Relationen zwischen der neuen, auf Momenten basierenden Distanz und anderen Distanzen auf Wahrscheinlichkeitsmaßen. Des Weiteren leiten wir mit Hilfe unseres Frameworks eine obere Schranke für das Missklassifikationsrisiko unserer Methode her. Um die Relevanz unseres theoretischen Frameworks zu untermauern, führen wir empirische Experimente auf zahlreichen großen Datenbanken durch. Die Resultate zeigen, dass unsere Methode, obwohl sie auf schwächeren Annahmen basiert, oft ähnliche alternative Methoden übertrifft, welche auf stärkeren Annahmen basieren.
Im dritten Teil wenden wir unser Framework auf zwei industrielle Regressionsprobleme an. Das erste Problem stammt aus dem Bereich der industriellen Produktion. Wir stellen einen neuen Algorithmus vor, der auf der Ähnlichkeit der ersten Momente von mehreren Wahrscheinlichkeitsverteilungen basiert. Unser Algorithmus ermöglicht die Modellierung von neuen, nicht der Wahrscheinlichkeitsverteilung der Trainingsdaten folgenden Zeitreihen und übertrifft, auf Datensätzen realer Problemstellungen, zahlreiche Standardregressionsalgorithmen. Das zweite Problem stammt aus dem Bereich der Analytischen Chemie. Wir stellen einen neuen, auf Momenten basierenden Algorithmus zur Kalibrierung chemischer Messsysteme vor. Im Gegensatz zu Standardalgorithmen basiert unser Algorithmus nur auf ungelabelten Daten des Anwendungsmesssystems. Wir diskutieren theoretische Eigenschaften des vorgestellten Algorithmus und zeigen, dass unser Algorithmus Standardalternativen oft übertrifft.This thesis contributes to the mathematical foundation of domain adaptation as emerging field in machine learning. In contrast to classical statistical learning, the framework of domain adaptation takes into account deviations between probability distributions in the training and application setting. Domain adaptation applies for a wider range of applications as future samples often follow a distribution that differs from the ones of the training samples. A decisive point is the generality of the assumptions about the similarity of the distributions. Therefore, in this thesis we study domain adaptation problems under as weak similarity assumptions as can be modelled by finitely many moments.
By examining the generalization ability of discriminative models trained under this relaxed assumption we establish, in the first part, a framework for bounding the misclassification risk based on finitely many moments and additional smoothness conditions. Our results show that a low misclassification risk of the discriminative models can be expected if a) the misclassification risk on the training sample is small, b) the sample size is large enough, and c) the samples' distributions meet an additional entropy condition.
In the second part, we apply our theoretical framework to the design of machine learning algorithms for domain adaptation. We propose a new moment distance for metric-based regularization of neural networks. Our methods aim at finding new data representations such that our weak assumptions on the similarity of the distributions are satisfied. In this context, various relations of the new moment distance to other probability metrics are proven. Further, a bound on the misclassification risk of our method is derived. To underpin the relevance of our theoretical framework, we perform empirical experiments on several large-scale benchmark datasets. The results show that our method, though based on weaker assumptions, often outperforms related alternatives based on stronger assumptions on the similarity of distributions.
In the third part, we apply our framework on two industrial regression problems. The first problem is settled in the area of industrial manufacturing. We propose a new algorithm that is based on the similarity of the first moments of multiple different distributions. Our algorithm enables the modeling of time series from previously unseen distributions and outperforms several standard regression algorithms on real-world data. The second problem stems from the area of analytical chemistry. We propose a new moment-based domain adaptation algorithm for the calibration of chemical measurement systems. In contrast to standard approaches, our algorithm is only based on unlabeled data from the application system. Theoretical properties of the proposed algorithm are discussed and it is shown to empirically outperform standard alternatives on two real-world datasets.submitted by DI Werner ZellingerAbweichender Titel laut Übersetzung der Verfasserin/des VerfassersDissertation Universität Linz 202
- …
