No description available.

No description available.

The purpose of this study is to investigate the medial portrayal of Romani people in the newspaper Dagens Nyheter from 2013 and 2014. In the theoretical chapter the concepts of new racism is used to speak of what is considered racist in the Swedish society today. Other concepts that are used are type and stereotype which are used to explain how our context is depending on distinctions to understand what surrounds us. Concepts such as myths and binary-oppositions are other theories that are used in this study.  Six debate articles have been selected to investigate how the Romani people are represented. Discourse analysis was used to answer the main questions in this study and was used to answer the purpose of this study. The main result was that the Romani people was discriminated and were not given the opportunity to speak their mind. It was always someone else outside this ethnic group who presented their view on something. They were considered passive as a group, in need of help from the remaining society. Said newspapers made generalizations about the group and connected them to begging. Finally, the perhaps most interesting result that this study finds is that the media makes distinctions between the Romani people and other social groups. The Romani people are made out as “the others” and a “we” and “them” in the Swedish society becomes prominent.

Modern deep neural networks are proficient at solving various visual recognition and understanding tasks, as long as a sufficiently large labeled dataset is available during the training time. However, the progress of these visual tasks is limited by the number of manual annotations. On the other hand, it is usually time-consuming and error-prone to annotate visual data, rendering the challenge of scaling up human labeling for many visual tasks. Fortunately, it is easy to collect large-scale, diverse unlabeled visual data from the Internet. And we can acquire a large amount of synthetic visual data with annotations from game engines effortlessly. In this dissertation, we explore how to utilize the unlabeled data and synthetic labeled data for various visual tasks, aiming to replace or reduce the direct supervision from the manual annotations. The key idea is to encourage deep neural networks to produce consistent predictions across different transformations (\eg geometry, temporal, photometric, etc.). We organize the dissertation as follows. In Part I, we propose to use the consistency over different geometric formulations and a cycle consistency over time to tackle the low-level scene geometry perception tasks in a self-supervised learning setting. In Part II, we tackle the high-level semantic understanding tasks in a semi-supervised learning setting, with the constraint that different augmented views of the same visual input maintain consistent semantic information. In Part III, we tackle the cross-domain image segmentation problem. By encouraging an adaptive segmentation model to output consistent results for a diverse set of strongly-augmented synthetic data, the model learns to perform test-time adaptation on unseen target domains with one single forward pass, without model training or optimization at the inference time. / Doctor of Philosophy / Recently, deep learning has emerged as one of the most powerful tools to solve various visual understanding tasks. However, the development of deep learning methods is significantly limited by the amount of manually labeled data. On the other hand, it is usually time-consuming and error-prone to annotate visual data, making the human labeling process not easily scalable. Fortunately, it is easy to collect large-scale, diverse raw visual data from the Internet (\eg search engines, YouTube, Instagram, etc.). And we can acquire a large amount of synthetic visual data with annotations from game engines effortlessly. In this dissertation, we explore how we can utilize the raw visual data and synthetic data for various visual tasks, aiming to replace or reduce the direct supervision from the manual annotations. The key idea behind this is to encourage deep neural networks to produce consistent predictions of the same visual input across different transformations (\eg geometry, temporal, photometric, etc.). We organize the dissertation as follows. In Part I, we propose using the consistency over different geometric formulations and a forward-backward cycle consistency over time to tackle the low-level scene geometry perception tasks, using unlabeled visual data only. In Part II, we tackle the high-level semantic understanding tasks using both a small amount of labeled data and a large amount of unlabeled data jointly, with the constraint that different augmented views of the same visual input maintain consistent semantic information. In Part III, we tackle the cross-domain image segmentation problem. By encouraging an adaptive segmentation model to output consistent results for a diverse set of strongly-augmented synthetic data, the model learns to perform test-time adaptation on unseen target domains.

This thesis addresses the problem of domain shift in 3D point cloud perception. In the last decades, there has been tremendous progress in within-domain training and testing. However, the performance of perception models is affected when training on a source domain and testing on a target domain sampled from different data distributions. As a result, a change in sensor or geo-location can lead to a harmful drop in model performance. While solutions exist for image perception, addressing this problem in point clouds remains unresolved. The focus of this thesis is the study and design of solutions for mitigating domain shift in 3D point cloud perception. We identify several settings differing in the level of target supervision and the availability of source data. We conduct a thorough study of each setting and introduce a new method to solve domain shift in each configuration. In particular, we study three novel settings in domain adaptation and domain generalization and propose five new methods for mitigating domain shift in 3D point cloud perception. Our methods are used by the research community, and at the time of writing, some of the proposed approaches hold the state-of-the-art. In conclusion, this thesis provides a valuable contribution to the computer vision community, setting the groundwork for the development of future works in cross-domain conditions.

This thesis explores the use of reinforcement learning within the Unity ML Agents framework to simulate zombie-like movements in humanoid ragdolls. The generated locomotion aims to embrace the Uncanny Valley phenomenon, partly through the way it walks, but also through limb disablement. Additionally, the paper strives to test the effectiveness of reinforcement learning as a valuable tool for generative adaptive locomotion. The research implements reward functions and addresses technical challenges. It lays a focus on adaptability through the limb disablement system. A user study comparing the reinforcement learning agent to Mixamo animations evaluates the effectiveness of simulating zombie-like movements as well as if the Uncanny Valley phenomenon was achieved. Results show that while the reinforcement learning agent may lack believability and uncanniness when compared to the Mixamo animation, it features a level of adaptability that is worth expanding upon. Given the inconclusive results, there is room for further research on the topic to achieve the Uncanny Valley effect and enhance zombie-like locomotion with reinforcement learning.

Federated Learning is an emerging privacy-preserving distributed machine learning approach to building a shared model by performing distributed training locally on participating devices (clients) and aggregating the local models into a global one. As this approach prevents data collection and aggregation, it helps in reducing associated privacy risks to a great extent. However, the data samples across all participating clients are usually not independent and identically distributed (non-i.i.d.), and Out of Distribution (OOD) generalization for the learned models can be poor. Besides this challenge, federated learning also remains vulnerable to various attacks on security wherein a few malicious participating entities work towards inserting backdoors, degrading the generated aggregated model as well as inferring the data owned by participating entities. In this work, we propose an approach for learning invariant (causal) features common to all participating clients in a federated learning setup and analyse empirically how it enhances the Out of Distribution (OOD) accuracy as well as the privacy of the final learned model. Although Federated Learning allows for participants to contribute their local data without revealing it, it faces issues in data security and in accurately paying participants for quality data contributions. In this report, we also propose an EOS Blockchain design and workflow to establish data security, a novel validation error based metric upon which we qualify gradient uploads for payment, and implement a small example of our Blockchain Causal Federated Learning model to analyze its performance with respect to robustness, privacy and fairness in incentivization. / L’apprentissage fédéré est une approche émergente d’apprentissage automatique distribué préservant la confidentialité pour créer un modèle partagé en effectuant une formation distribuée localement sur les appareils participants (clients) et en agrégeant les modèles locaux en un modèle global. Comme cette approche empêche la collecte et l’agrégation de données, elle contribue à réduire dans une large mesure les risques associés à la vie privée. Cependant, les échantillons de données de tous les clients participants sont généralement pas indépendante et distribuée de manière identique (non-i.i.d.), et la généralisation hors distribution (OOD) pour les modèles appris peut être médiocre. Outre ce défi, l’apprentissage fédéré reste également vulnérable à diverses attaques contre la sécurité dans lesquelles quelques entités participantes malveillantes s’efforcent d’insérer des portes dérobées, dégradant le modèle agrégé généré ainsi que d’inférer les données détenues par les entités participantes. Dans cet article, nous proposons une approche pour l’apprentissage des caractéristiques invariantes (causales) communes à tous les clients participants dans une configuration d’apprentissage fédérée et analysons empiriquement comment elle améliore la précision hors distribution (OOD) ainsi que la confidentialité du modèle appris final. Bien que l’apprentissage fédéré permette aux participants de contribuer leurs données locales sans les révéler, il se heurte à des problèmes de sécurité des données et de paiement précis des participants pour des contributions de données de qualité. Dans ce rapport, nous proposons également une conception et un flux de travail EOS Blockchain pour établir la sécurité des données, une nouvelle métrique basée sur les erreurs de validation sur laquelle nous qualifions les téléchargements de gradient pour le paiement, et implémentons un petit exemple de notre modèle d’apprentissage fédéré blockchain pour analyser ses performances.

Malgré des progrès remarquables dans une grande variété de sujets, les réseaux de neurones éprouvent toujours des difficultés à exécuter certaines tâches simples pour lesquelles les humains excellent. Comme indiqué dans des travaux récents, nous émettons l'hypothèse que l'écart qualitatif entre l'apprentissage en profondeur actuel et l'intelligence humaine est le résultat de biais inductifs essentiels manquants. En d'autres termes, en identifiant certains de ces biais inductifs essentiels, nous améliorerons le transfert d'informations dans les réseaux artificiels, ainsi que certaines de leurs limitations actuelles les plus importantes sur un grand ensemble de tâches. Les limites sur lesquelles nous nous concentrerons dans cette thèse sont la généralisation systématique hors distribution et la capacité d'apprendre sur des échelles de temps extrêmement longues. Dans le premier article, nous nous concentrerons sur l'extension des réseaux de neurones récurrents (RNN) à contraintes spectrales et proposerons une nouvelle structure de connectivité basée sur la décomposition de Schur, en conservant les avantages de stabilité et la vitesse d'entraînement des RNN orthogonaux tout en améliorant l'expressivité pour les calculs complexes à court terme par des dynamiques transientes. Cela sert de première étape pour atténuer le problème du "exploding vanishing gradient" (EVGP). Dans le deuxième article, nous nous concentrerons sur les RNN avec une mémoire externe et un mécanisme d'auto-attention comme un moyen alternatif de résoudre le problème du EVGP. Ici, la contribution principale sera une analyse formelle sur la stabilité asymptotique du gradient, et nous identifierons la pertinence d'événements comme un ingrédient clé pour mettre à l'échelle les systèmes d'attention. Nous exploitons ensuite ces résultats théoriques pour fournir un nouveau mécanisme de dépistage de la pertinence, qui permet de concentrer l'auto-attention ainsi que de la mettre à l'échelle, tout en maintenant une bonne propagation du gradient sur de longues séquences. Enfin, dans le troisième article, nous distillons un ensemble minimal de biais inductifs pour les tâches cognitives purement relationnelles et identifions que la séparation des informations relationnelles des entrées sensorielles est un ingrédient inductif clé pour la généralisation OoD sur des entrées invisibles. Nous discutons en outre des extensions aux relations non-vues ainsi que des entrées avec des signaux parasites. / Despite remarkable advances in a wide variety of subjects, neural networks are still struggling on simple tasks humans excel at. As outlined in recent work, we hypothesize that the qualitative gap between current deep learning and human-level artificial intelligence is the result of missing essential inductive biases. In other words, by identifying some of these key inductive biases, we will improve information transfer in artificial networks, as well as improve on some of their current most important limitations on a wide range of tasks. The limitations we will focus on in this thesis are out-of-distribution systematic generalization and the ability to learn over extremely long-time scales. In the First Article, we will focus on extending spectrally constrained Recurrent Neural Networks (RNNs), and propose a novel connectivity structure based on the Schur decomposition, retaining the stability advantages and training speed of orthogonal RNNs while enhancing expressivity for short-term complex computations via transient dynamics. This serves as a first step in mitigating the Exploding Vanishing Gradient Problem (EVGP). In the Second Article, we will focus on memory augmented self-attention RNNs as an alternative way to tackling the Exploding Vanishing Gradient Problem (EVGP). Here the main contribution will be a formal analysis on asymptotic gradient stability, and we will identify event relevancy as a key ingredient to scale attention systems. We then leverage these theoretical results to provide a novel relevancy screening mechanism, which makes self-attention sparse and scalable, while maintaining good gradient propagation over long sequences. Finally, in the Third Article, we distill a minimal set of inductive biases for purely relational cognitive tasks, and identify that separating relational information from sensory input is a key inductive ingredient for OoD generalization on unseen inputs. We further discuss extensions to unseen relations as well as settings with spurious features.

L'intelligence artificielle est un domaine en pleine évolution. Au premier plan des percées récentes se retrouve des approches connues sous le nom d'apprentissage automatique. Cependant, bien que l'apprentissage automatique ait montré des performances remarquables dans des tâches telles que la reconnaissance et la génération d'images, la génération et la traduction de textes et le traitement de la parole, il est connu pour échouer silencieusement dans des conditions courantes. Cela est dû au fait que les algorithmes modernes héritent des biais des données utilisées pour les créer, ce qui conduit à des prédictions incorrectes lorsqu'ils rencontrent de nouvelles données différentes des données d'entraînement. Ce problème est connu sous le nom de défaillance hors-distribution. Cela rend l'intelligence artificielle moderne peu fiable et constitue un obstacle important à son déploiement sécuritaire et généralisé. Ignorer l'échec de généralisation hors-distribution de l'apprentissage automatique pourrait entraîner des situations mettant des vies en danger. Cette thèse vise à aborder cette question et propose des solutions pour assurer le déploiement sûr et fiable de modèles d'intelligence artificielle modernes. Nous présentons trois articles qui couvrent différentes directions pour résoudre l'échec de généralisation hors-distribution de l'apprentissage automatique. Le premier article propose une approche directe qui démontre une performance améliorée par rapport à l'état de l'art. Le deuxième article établie les bases de recherches futures en généralisation hors distribution dans les séries temporelles, tandis que le troisième article fournit une solution simple pour corriger les échecs de généralisation des grands modèles pré-entraînés lorsqu'entraîné sur tes tâches en aval. Ces articles apportent des contributions précieuses au domaine et fournissent des pistes prometteuses pour la recherche future en généralisation hors distribution. / Artificial Intelligence (AI) is a rapidly advancing field, with data-driven approaches known as machine learning, at the forefront of many recent breakthroughs. However, while machine learning have shown remarkable performance in tasks such as image recognition and generation, text generation and translation, and speech processing, they are known to silently fail under common conditions. This is because modern AI algorithms inherit biases from the data used to train them, leading to incorrect predictions when encountering new data that is different from the training data. This problem is known as distribution shift or out-of-distribution (OOD) failure. This causes modern AI to be untrustworthy and is a significant barrier to the safe widespread deployment of AI. Failing to address the OOD generalization failure of machine learning could result in situations that put lives in danger or make it impossible to deploy AI in any significant manner. This thesis aims to tackle this issue and proposes solutions to ensure the safe and reliable deployment of modern deep learning models. We present three papers that cover different directions in solving the OOD generalization failure of machine learning. The first paper proposes a direct approach that demonstrates improved performance over the state-of-the-art. The second paper lays the groundwork for future research in OOD generalization in time series, while the third paper provides a straightforward solution for fixing generalization failures of large pretrained models when finetuned on downstream tasks. These papers make valuable contributions to the field and provide promising avenues for future research in OOD generalization.

Made available in DSpace on 2016-04-29T13:18:02Z (GMT). No. of bitstreams: 1 Reneta S H Pasquinelli.pdf: 1572588 bytes, checksum: af604fc1ca48edd37db7337c257770a5 (MD5) Previous issue date: 2007-05-03 / The present study s goals were to verify (1) the control of distinct audiences over the theme of spoken verbal behavior; (2) the generalization of audience control to new audiences, with distinct physical features; (3) how much direct reinforcement of a given repertoire was necessary for a new audience to assume evocative control over the repertoire; (4) how much direct reinforcement was necessary in the presence of a new audience, to establish the control of this new audience over a second repertoire. Six children, ranging from 4 to 8 years participated in the study. Four puppets, 2 of them humanlike were established as audiences during the experiment. On each condition 2 puppets with distinct features asked each participant to describe 5s films. Each film portrayed a person or child engaged in some action. On the first 3 experimental conditions the first pair of puppets trained and/or tested the emergence of 2 verbal repertoires descriptive of the films: a mentalist/ internalist repertoire composed of descriptions of supposed emotions or purposes (m repertoire) and a externalist repertoire with descriptions of actions or physical characteristics (p repertoire). On the last conditions a second pair of puppets was used to test if direct reinforcement of some responses belonging to one of the repertoires would evoke the same repertoire on new trials with new films. Tests of the effects of a reversal condition with new audiences were also conducted. Results showed (1) the establishment of the puppets as audiences controlling different thematic repertoires; (2) the occurrence of generalization of this evocative function to a new audience after the direct reinforcement of a few responses; (3) the reversal of repertoires evoked by an audience after the reinforcement of a few responses belonging to this repertoire; (4) the emergence of variability on the verbal repertoire of the participants during tests. The role of the audience as discriminative and conditional stimuli and the variability of the verbal responses are discussed / O presente estudo pretendeu verificar se: (1) é possível estabelecer o controle de duas audiências distintas sobre o repertório verbal de participantes com a função de selecionar o assunto que se fala? (2) Há generalização do controle de uma audiência para uma nova audiência, com características físicas distintas, sendo suficiente que sejam reforçadas apenas algumas das respostas previamente evocadas na presença da primeira audiência? (3) Quanto de um dado repertório deveria ser reforçado para que uma nova audiência passe a ter função evocativa sobre um determinado repertório verbal? (4) Quantas respostas precisam ser reforçadas para que uma audiência que já controla um repertório verbal, passe a controlar um segundo repertório? Participaram do estudo 6 crianças, de 4 a 8 anos. Foram utilizados 4 fantoches, 2 com características humanas e 2 com características animais. Em cada fase 2 fantoches, com características físicas diversas, requisitavam que o participante descrevesse filmes de 5s, que continham uma pessoa realizando alguma atividade. Foram estabelecidos 2 repertórios descritivos dos filmes: repertório f (descrições de características físicas e de ações dos personagens) e m (descrições que supunham finalidade e emoções aos personagens). Nas 3 primeiras fases uma dupla de fantoches treinou e testou o estabelecimento dos repertórios f e m. Posteriormente, outros 2 fantoches testaram se o reforçamento de algumas das respostas pertencentes a um dos repertórios estabelecidos promoveria generalização do controle para as novas audiências. Também foram testados os efeitos de uma condição de reversão sobre as nova audiências. Os resultados de todos os participantes indicaram: o estabelecimento dos fantoches como audiências que controlam diferentes temas / assuntos do repertório do falante; (2) a ocorrência da generalização desta função evocativa para uma nova audiência pelo reforçamento de algumas respostas do repertório controlado por uma outra audiência; (3) a reversão dos repertórios evocados pelas audiências depois do reforçamento de algumas respostas no treino com reversão; (4) a ocorrência de variabilidade nas respostas dos participantes no decorrer do experimento. Discute-se o papel da audiência como estímulo discriminativo ou condicional e a variabilidade do comportamento verbal