Modern DBMSes are designed to support many transactions running simultaneously. DBMS thrashing is indicated by the existence of a sharp drop in transaction throughput. The thrashing behavior in DBMSes is a serious concern to DBAs engaged in on-line transaction processing (OLTP) and on-line analytical processing (OLAP) systems, as well as to DBMS implementors developing technologies related to concurrency control. If thrashing is prevalent in a DBMS, thousands of transactions may be aborted, resulting in little progress in transaction throughput over time. From an engineering perspective, therefore, it is of critical importance to understand the factors of DBMS thrashing. However, understanding the origin of modern DBMSes' thrashing is challenging, due to many factors that may interact. The existing literature on thrashing exhibits the following weaknesses: (i) methodologies have been based on simulation and analytical studies, rather than on empirical analysis on real DBMSes, (ii) scant attention has been paid to the associations between factors, and (iii) studies have been restricted to one specific DBMS rather than across multiple DBMSes. This dissertation aims at better understanding the thrashing phenomenon across multiple DBMSes. We identify the underlying causes and propose a novel structural causal model to explicate the relationships between various factors contributing to DBMS thrashing. Our model derives a number of specific hypotheses to be subsequently tested across DBMSes, providing empirical support for this model as well as engineering implications for fundamental improvements in transaction processing. Our model also guides database researchers to refine this causal model, by looking into other unknown factors.

<p dir="ltr">Recent advancements in representation learning have significantly enhanced the analysis of relational data across various domains, including social networks, bioinformatics, and recommendation systems. In general, these methods assume that the training and test datasets come from the same distribution, an assumption that often fails in real-world scenarios due to evolving data, privacy constraints, and limited resources. The task of out-of-distribution (OOD) extrapolation emerges when the distribution of test data differs from that of the training data, presenting a significant, yet unresolved challenge within the field. This dissertation focuses on developing robust representations for effective OOD extrapolation, specifically targeting relational data types like graphs and sets. For successful OOD extrapolation, it's essential to first acquire a representation that is adequately expressive for tasks within the distribution. In the first work, we introduce Set Twister, a permutation-invariant set representation that generalizes and enhances the theoretical expressiveness of DeepSets, a simple and widely used permutation-invariant representation for set data, allowing it to capture higher-order dependencies. We showcase its implementation simplicity and computational efficiency, as well as its competitive performances with more complex state-of-the-art graph representations in several graph node classification tasks. Secondly, we address OOD scenarios in graph classification and link prediction tasks, particularly when faced with varying graph sizes. Under causal model assumptions, we derive approximately invariant graph representations that improve extrapolation in OOD graph classification task. Furthermore, we provide the first theoretical study of the capability of graph neural networks for inductive OOD link prediction and present a novel representation model that produces structural pairwise embeddings, maintaining predictive accuracy for OOD link prediction as the test graph size increases. Finally, we investigate the impact of environmental data as a confounder between input and target variables, proposing a novel approach utilizing an auxiliary dataset to mitigate distribution shifts. This comprehensive study not only advances our understanding of representation learning in OOD contexts but also highlights potential pathways for future research in enhancing model robustness across diverse applications.</p>

L'intérêt en apprentissage machine pour étudier la causalité s'est considérablement accru ces dernières années. Cette approche est cependant encore peu répandue dans le domaine de l’intelligence artificielle (IA) et du droit. Elle devrait l'être. L'approche associative actuelle d’apprentissage machine révèle certaines limites que l'analyse causale peut surmonter. Cette thèse vise à découvrir si les modèles causaux peuvent être utilisés en IA et droit. Nous procédons à une brève revue sur le raisonnement et la causalité en science et en droit. Traditionnellement, les cadres normatifs du raisonnement étaient la logique et la rationalité, mais la théorie duale démontre que la prise de décision humaine dépend de nombreux facteurs qui défient la rationalité. À ce titre, des statistiques et des probabilités étaient nécessaires pour améliorer la prédiction des résultats décisionnels. En droit, les cadres de causalité ont été définis par des décisions historiques, mais la plupart des modèles d’aujourd’hui de l'IA et droit n'impliquent pas d'analyse causale. Nous fournissons un bref résumé de ces modèles, puis appliquons le langage structurel de Judea Pearl et les définitions Halpern-Pearl de la causalité pour modéliser quelques décisions juridiques canadiennes qui impliquent la causalité. Les résultats suggèrent qu'il est non seulement possible d'utiliser des modèles de causalité formels pour décrire les décisions juridiques, mais également utile car un schéma uniforme élimine l'ambiguïté. De plus, les cadres de causalité sont utiles pour promouvoir la responsabilisation et minimiser les biais. / The machine learning community’s interest in causality has significantly increased in recent years. This trend has not yet been made popular in AI & Law. It should be because the current associative ML approach reveals certain limitations that causal analysis may overcome. This research paper aims to discover whether formal causal frameworks can be used in AI & Law. We proceed with a brief account of scholarship on reasoning and causality in science and in law. Traditionally, normative frameworks for reasoning have been logic and rationality, but the dual theory has shown that human decision-making depends on many factors that defy rationality. As such, statistics and probability were called for to improve the prediction of decisional outcomes. In law, causal frameworks have been defined by landmark decisions but most of the AI & Law models today do not involve causal analysis. We provide a brief summary of these models and then attempt to apply Judea Pearl’s structural language and the Halpern-Pearl definitions of actual causality to model a few Canadian legal decisions that involve causality. Results suggest that it is not only possible to use formal causal models to describe legal decisions, but also useful because a uniform schema eliminates ambiguity. Also, causal frameworks are helpful in promoting accountability and minimizing biases.