Direct, construct by construct translation from one high level language to another often produces convoluted, unnatural, and unreadable results, particularly when the source and target languages support different models of programming. A more readable and natural translation can be obtained by augmenting the translator with a program transformation system.

Cette thèse s'intéresse à la métathéorie de la théorie des types intuitionniste. Les systèmes que nous considérons sont des variantes de la théorie des types de Martin-Löf ou du Calcul des Constructions, et nous nous intéressons à la cohérence de ces systèmes ou encore à l'indépendance d'axiomes par rapport à ces systèmes. Le fil rouge de cette thèse est la construction de modèles syntaxiques, qui sont des modèles qui réutilisent la théorie des types pour interpréter la théorie des types. Dans une première partie, nous introduisons la théorie des types à l'aide d'un système minimal et de plusieurs extensions potentielles. Dans une seconde partie, nous introduisons les modèles syntaxiques donnés par traduction de programme et donnons plusieurs exemples. Dans une troisième partie, nous présentons Template-Coq, un plugin de métaprogrammation pour Coq. Nous montrons comment l'utiliser pour implémenter directement certains modèles syntaxiques. Enfin, dans une dernière partie, nous nous intéressons aux théories des types à deux égalités : une égalité stricte et une égalité univalente. Nous proposons une relecture des travaux de Coquand et. al. et Orton et Pitts sur le modèle cubique en introduisant la notion de fibrance dégénérée. / This thesis is about the metatheory of intuitionnistic type theory. The considered systems are variants of Martin-Löf type theory of Calculus of Constructions, and we are interested in the coherence of those systems and in the independence of axioms with respect to those systems. The common theme of this thesis is the construction of syntactic models, which are models reusing type theory to interpret type theory. In a first part, we introduce type theory by a minimal system and several possible extensions. In a second part, we introduce the syntactic models given by program translation and give several examples. In a third part, we present Template-Coq, a plugin for metaprogramming in Coq. We demonstrate how to use it to implement directly some syntactic models. Last, we consider type theories with two equalities: one strict and one univalent. We propose a re-reading of works of Coquand et.al. and of Orton and Pitts on the cubical model by introducing degenerate fibrancy.

In recent years, deep learning based sequence modeling (neural sequence modeling) techniques have made substantial progress in many tasks, including information retrieval, question answering, information extraction, machine translation, etc. Benefiting from the highly scalable attention-based Transformer architecture and enormous open access online data, large-scale pre-trained language models have shown great modeling and generalization capacity for sequential data. However, not all domains benefit equally from the rapid development of neural sequence modeling. Domains like healthcare and software engineering have vast amounts of sequential data containing rich knowledge, yet remain under-explored due to a number of challenges: 1) the distribution of the sequences in specific domains is different from the general domain; 2) the effective comprehension of domain-specific data usually relies on domain knowledge; and 3) the labelled data is usually scarce and expensive to get in domain-specific settings. In this thesis, we focus on the research problem of applying neural sequence modeling methods to address both common and domain-specific challenges from the healthcare and software engineering domains. We systematically investigate neural-based machine learning approaches to address the above challenges in three research directions: 1) learning with long sequences, 2) learning from domain knowledge and 3) learning under limited supervision. Our work can also potentially benefit more domains with large amounts of sequential data. / Doctor of Philosophy / In the last few years, computer programs that learn and understand human languages (an area called machine learning for natural language processing) have significantly improved. These advances are visible in various areas such as retrieving information, answering questions, extracting key details from texts, and translating between languages. A key to these successes has been the use of a type of neural network structure known as a "Transformer", which can process and learn from lots of information found online. However, these successes are not uniform across all areas. Two fields, healthcare and software engineering, still present unique challenges despite having a wealth of information. Some of these challenges include the different types of information in these fields, the need for specific expertise to understand this information, and the shortage of labeled data, which is crucial for training machine learning models. In this thesis, we focus on the use of machine learning for natural language processing methods to solve these challenges in the healthcare and software engineering fields. Our research investigates learning with long documents, learning from domain-specific expertise, and learning when there's a shortage of labeled data. The insights and techniques from our work could potentially be applied to other fields that also have a lot of sequential data.