Contributions in Merlin

The Internet of Things (IoT) platform is one of the key drivers of the smart home market, having revolutionized the advancement of smart home technology. Besides the many benefits for convenience and efficiency, there are also concerns about security and privacy in such environments. The increasing complexity of smart homes and hardware limitations of individual devices necessitate the storage and processing of data in remote cloud environments. This raises privacy issues due to potential misuse or disclosure of sensitive information about residents. To the author's knowledge, no existing Privacy Enhancing Technology (PET) offers a lightweight approach to enforce privacy in smart home environments by combining existing tools into a unifying framework. The goal of this thesis is to take a first step towards an extensible open source software system that integrates into the smart home environment with the purpose of monitoring smart home device communications and controlling their communication behavior through user-defined policies. To this end, a prototype application is developed, which monitors smart home devices' Domain Name System (DNS) requests and enforces policies via a DNS sinkhole mechanism. The prototype system is deployed to a system-on-chip platform and evaluated in a live smart home environment to gain insight into the viability of the prototype. The aim is to examine the performance, effectiveness, and limitations of the prototype with the intention of validating the general approach. The results of these experiments indicate that the prototype successfully achieves the goals outlined in this thesis. The application prototype is capable of monitoring the network activity of smart home devices. The collected data are processed to gain insights and make this information transparent to the users. Furthermore, the prototype allows users to define simple allow/block policies which are subsequently enforced by the system.

The task of text simplification is to reduce the linguistic complexity of a text in order to make it more accessible. Previous work on text simplification has primarily focused on either a single level of simplification or multiple levels of simplification, but always with the goal of making the text simpler. In this work, we explore a related task: re-generating sentences to produce equivalent text that targets an audience at a different readability level, whether that level is simpler or more advanced. We formulate the problem as a sequence to sequence task and explore different methods of using the pre-trained T5 encoder-decoder model to perform the task. In particular, we investigate the use of the hyperformer++ \cite{mahabadi2021parameter} architecture to solve the task, and propose and evaluate custom variants of the architecture designed to maximize positive transfer between different transformation pairs. According to automatic metrics, our custom variant of hyperformer++ is able to compete with strong baselines while only storing a small fraction of parameters compared to updating the entire language model.

Histograms are widely used because they are easy to implement and provide a simple overview of the underlying data. However, histograms are limited to two dimensions and thus not suited for multi-dimensional data. To resolve this, several models have been designed in the existing literature. These typically combine parallel coordinates plot (PCP) with histograms, so that they can represent multidimensional data. However, these existing models typically do not enable clustering of multivariate data or user interaction. To fill this gap, this thesis introduces a new "clustering PHP application" which offers a visual explorative framework with user interaction for the purpose of clustering. This application integrates PHP, Principal Component analysis (PCA), and scatter plots to merge their respective advantages. First, the PCA part offers ideas about variables such as how important they are and how they are related. Variables of interest can then be plotted on the PHP, which was adjusted for clustering (clustering PHP), to visually find relationships between variables. Axes on the clustering PHP can be reordered to focus on specific variables. Finally, a scatter plot helps users to observe local features and allows for the selection of principal components or variables. Interactions are immediately synchronized on the scatter plot and clustering PHP to detect data points sharing similarities on subspaces effortlessly. Overall this "clustering PHP application" thus helps users to determine clustering groups and improve clustering accuracy. In summary, "clustering PHP application" can help a user to explore data and make subspace clustering with complex multi-dimensional data more easy and efficient.

Die Förderung von ethischem Verhalten hat im Zusammenhang mit Fehlverhalten und Unternehmens- skandalen immer mehr an Bedeutung gewonnen. Diese Arbeit hat das Ziel, in diesem Rahmen die Forschungsfrage zu beantworten, ob Contemplation Questions und Regulatory Focus die Entscheidungs- fähigkeit unterstützen können. Dazu wurde eine Literaturrecherche zur Vertiefung des Verständnisses durchgeführt und darauf aufbauend wurden Annahmen und Hypothesen gebildet, die mittels Gedankenexperiment untersucht wurden. Die Erkenntnisse führen zum Schluss, dass Contemplation Questions und Regulatory Focus vereinzelt in der Lage sein sollten, die Entscheidungsfähigkeit unter passenden Bedingungen zu unterstützen. Ob dies jedoch in Verbindung der beiden Komponenten möglich ist, kann abschliessend nur durch weiterführende empirische Untersuchungen bestätigt werden.

Deep convolutional neural networks (CNNs) have become the state-of-the-art approach for image classification. While these networks are very effective at identifying the class to which an image belongs, they often do not properly learn the semantic relationship between classes. This means that models treat all misclassifications equally during training, regardless of the semantic distance between the predicted and actual class. This approach does not reflect the complexity of the real world, where some entities are more similar to each other, making mistakes between related classes less severe than those of unrelated classes. An architecture suited for hierarchical classification is presented as a potential solution to this problem. Rather than just predicting a single class, networks predict a simplified hierarchy consisting of higher-level concepts. This thesis explores how the architecture of CNNs can be adapted to incorporate hierarchical information to increase performance and the semantical conditioning of CNNs. The ultimate goal is to enhance the accuracy and robustness of image classification models by improving their understanding of the semantic relationships between classes, which could potentially lead to fewer and less severe misclassifications. To achieve this, several architectures are explored -- all using a ResNet backbone with classifiers for each hierarchical level -- that are compared with a baseline model that does not utilize the hierarchy for predictions. Most importantly, this thesis proposes an attention mechanism that does not contain any extra trainable parameters. This attention mechanism transforms the deep features given to a lower-level classifier based on the weight matrix from the higher-level classifier. This transformation aims to highlight features relevant to the classification of the higher-level concept, thus enabling the model to learn the decision boundary between classes of different higher-level concepts. This attention mechanism can effectively increase the classification accuracy for the ImageNet classes compared to a baseline architecture. Furthermore, when provided ground-truth information about the hierarchies from classes during training, it effectively learns the decision boundaries between classes from different higher-level concepts. This thesis also explores whether these architectures can be used for open-set classification. While showing some potential, the attention mechanism could likely be adapted for open-set classification, representing a promising possibility for future research.

Open-Set classification (OSC) addresses one of the core issues of traditional classification techniques, namely, the underlying closed-world assumption. The goal of OSC methods is to classify known classes correctly while also rejecting unknown classes. We propose two novel generic loss functions, Margin-OS and Margin-EOS, which combine the Entropic Open-Set and Objectosphere loss with margin-based loss functions used in face recognition tasks, CosFace and ArcFace, to learn discriminative features. We find that the margin has a positive effect on the closed-set accuracy but a mixed effect on the open-set performance. For applications that can tolerate high false positive rates, our losses improve the classification of known classes, but for low false positive rates the margin negatively impacts the training which leads to subpar classification of known samples.

Ein gut funktionierender Elektrizitätsmarkt ist von essenzieller Bedeutung für unserer Gesellschaft. Um diesen besser zu verstehen, wurde er bereits umfassend erforscht. Die Coronapandemie und Ukrainekrise haben jedoch grosse Veränderungen mit sich gebracht, die bis jetzt kaum untersucht wurden. Vorliegende Arbeit erleuchtet diese Forschungslücke indem Strompreise und -verbräuche mittels Korrelationen und Regressionen analysiert werden. Die Ergebnisse zeigen, dass die Elektrizitätsmärkte während den Preisexplosionen in den Jahren 2021 und 2022 stärker voneinander abhängen und die Nachfrage nach Elektrizität die Preise nahezu nicht beeinflusst hat. In der Schweiz konnte im Jahr 2022 sogar eine leicht negative Korrelation zwischen Preis und Verbrauch festgestellt werden.

Value-at-Risk (VaR) is a widely used statistical measure of financial risk. It provides an estimate of the maximum expected loss at a given confidence level over a specified time horizon. VaR models can be based on different statistical approaches, including Parametric models (Standard Normal, Weighted Standard Normal Increasing and Decreasing) and NonParametric models (Historical, Weighted Historical Increasing and Decreasing). However, there is no consensus on which approach is the best for predicting financial risk, and the choice of model can have a significant impact on the accuracy and robustness of VaR estimates. The goal of this study is to compare the performance of two different VaR models in forecasting the Value at Risk of different indices from the three asset classes of Equities, Commodities, and Fixed Income. The models to be tested are the Standard Normal and Historical, and for both we will analyze the three cases of Equally Weighted, Weighted Increasing and Weighted Decreasing. The primary objective of this study is to identify the model that provides the most accurate and robust predictions of risk for the three different indices. To achieve this objective, we will calculate VaR using the two different models for three different indices from three different asset classes. For each of the two models, we will apply different weights to the sample observations, equal, increasing and decreasing, so that each time different importance will be attributed to more dated or recent data. We will then backtest the VaR estimates using the Traffic Light Approach from the Basel II regulation, which is a supervisory tool used by regulators to assess the operational risk management practices of banks. We will classify the VaR estimates into three categories based on their performance: green, yellow, and red. The green category represents VaR estimates that perform well, the yellow category represents VaR estimates that need improvement, and the red category represents VaR estimates that are not acceptable. We will compare the performance of the VaR models based on the number of green lights achieved during the backtesting process. We will also analyze the results to determine which VaR model is the most robust and accurate for predicting the risk of different indices. Moreover, we will compare the results obtained utilizing 2-years data with the ones obtained utilizing 10-years data in order to add robustness to the findings. The expected findings of this study are that one of the VaR models will perform better than the others in terms of accuracy and robustness. We also expect to observe that for different asset classes the best performing model will vary, showing how one or another model best suits the different characteristics of each. The findings of this study will contribute to the existing literature on VaR modeling and model selection. In conclusion, this study aims to provide insights into the performance of different VaR models and their suitability for predicting financial risk. The findings of this study will be of interest to risk managers, investors, and regulators who use VaR as a tool for measuring and managing financial risk.

In dieser Arbeit wird mit dem Fama-French-3-Factor- und Carhart-4-Factor-Model eine empirische Analyse im Schweizer Aktienmarkt durchgeführt. Im untersuchten Zeitraum von Juli 2012 bis Juni 2022 werden die Risk Premiums dieser Modelle berechnet. Unabhängig von den gewählten Modellannahmen können das durchschnittliche Market Risk und Momentum Premium signifikant ungleich null nachgewiesen werden. Das mittlere Size Premium wird Annahmen-unabhängig nicht signifikant ungleich null belegt. Die Schlussfolgerung zur Signifikanz des durchschnittlichen Value Premiums fällt aufgrund der Abhängigkeit zu den gewählten Modellannahmen nicht eindeutig aus. Der Modellvergleich mittels Regressionsanalyse zeigt, dass das Carhart-Model, auch unter adjustierten Bedingungen, durchgehend die höhere Modellgüte als das Fama-French-Model aufweist

I conduct a longitude study of the carry trade with 2008 as the cutting point. Employing a sample with 17 currencies and over 30 years since 1993, I find that the basic carry trade strategy with G10 currencies does underperform after 2008 relative to itself before 2008 and relative to the market. I demonstrate that refined carry strategies including basic carry with an extended currency basket, carry momentum, and volatility-adjusted carry can avoid this downturn. I observe that overall traditional risk factors fail to explain the excess return of the carry trade. However, I show that during certain periods more significance can be observed, and higher returns are associated with a lower and more negative correlation with the market. Keywords: Carry, Momentum, Investment Strategy, Factor Model