Video action understanding involves comprehending actions performed by humans, depicted in videos. Central to the task of video action understanding are four fundamental questions: What, When, Where, and Who. These questions encapsulate the essence of action classification, temporal action localization, action detection, and actor recognition. Despite notable progress in research related to these tasks, many challenges persist and in this dissertation, we propose innovative solutions to tackle these challenges head-on. First, we address the challenges in action classification (``What?"), specifically related to multi-view action recognition. We propose a novel transformer decoder-based model, with learnable view and action queries, to enforce the learning of action features robust to shifts in viewpoints. Next, we focus on temporal action localization (``What?" and ``When?") and address challenges introduced in the multi-label setting. Our proposed solution involves leveraging the inherent relationships between complex actions in real-world videos. We introduce an attention-based architecture that models these relationships for the task of temporal action localization. Next, we propose \textit{Gabriella}, a real-time online system for activity detection (``What?", ``When?", and ``Where?") in security videos. Our proposed solution has three stages: tubelet extraction, activity classification, and online tubelet merging. For tubelet extraction, we propose a localization network that detects potential foreground regions to generate action tubelets. The detected tubelets are assigned activity class scores by the classification network and merged using our proposed Tubelet-Merge Action-Split (TMAS) algorithm to form the final action detections. Finally, we introduce an approach to solve the novel task of joint action and actor recognition (``What?" and ``Who?") and solve it using disentangled representation learning. We introduce a novel method to simultaneously identify both subjects (actors) and their actions. Our transformer-based model learns to separate actor and action features effectively by employing supervised contrastive losses alongside standard cross-entropy loss to ensure proper feature separation.

No / The UK government’s Equality Bill is rapidly changing the approach to equality in the workplace. This extends previous legislation which, in addition to requiring public authorities to promote equal opportunities, provides support for specific types of positive action. However, there continues to be a level of confusion, misinterpretation and sometimes suspicion of positive action. In general, this has resulted in an apparent reluctance to use positive action to promote representational diversity in the workplace. This is compounded by limited empirical evidence on its use. NHS Employers (the employers’ organisation for the National Health Service) in partnership with Bradford University Centre for Diversity and Inclusion collaborated on a project to establish where positive action is being used in the NHS and the key factors associated with its successful application. The work involved 20 NHS organisations in England and employed a mixed method of qualitative research including in-depth interviews, a mapping exercise and focus groups. This paper describes the methodology adopted in the study and some of the positive action initiatives employed. Key factors for making positive action a success are identified and recommendations are made which will have relevance for future policy direction.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

The goal of this thesis is to develop a set of tools for continuous tracking of behavioral phenomena in videos to support human behavior study. Current standard practices for extracting useful behavioral information from a video are typically difficult to replicate and require a lot of human time. For example, extensive training is typically required for a human coder to reliably code a particular behavior/interaction. Also, manual coding typically takes a lot more time than the actual length of the video (e.g. , it can take up to 6 times the actual length of the video to do human-assisted single object tracking. The time intensive nature of this process (due to the need to train expert and manual coding) puts a strong burden on the research process. In fact, it is not uncommon for an institution that heavily uses videos for behavioral research to have a massive backlog of unprocessed video data. To address this issue, I have developed an efficient behavior retrieval and interactive tracking system. These tools allow behavioral researchers/clinicians to more easily extract relevant behavioral information, and more objectively analyze behavioral data from videos. I have demonstrated that my behavior retrieval system achieves state-of-the-art performance for retrieving stereotypical behaviors of individuals with autism in a real-world video data captured in a classroom setting. I have also demonstrated that my interactive tracking system is able to produce high-precision tracking results with less human effort compared to the state-of-the-art. I further show that by leveraging the tracking results, we can extract an objective measure based on proximity between people that is useful for analyzing certain social interactions. I validated this new measure by showing that we can use it to predict qualitative expert ratings in the Strange Situation (a procedure for studying infant attachment security), a quantity that is difficult to obtain due to the difficulty in training the human expert.

Ce mémoire porte sur quelques notions appropriées d'actions de groupe sur les variétés symplectiques, à savoir en ordre décroissant de généralité : les actions symplectiques, les actions faiblement hamiltoniennes et les actions hamiltoniennes. Une connaissance des actions de groupes et de la géométrie symplectique étant prérequise, deux chapitres sont consacrés à des présentations élémentaires de ces sujets. Le cas des actions hamiltoniennes est étudié en détail au quatrième chapitre : l'importante application moment y est définie et plusieurs résultats concernant les orbites de la représentation coadjointe, tels que les théorèmes de Kirillov et de Kostant-Souriau, y sont démontrés. Le dernier chapitre se concentre sur les actions hamiltoniennes des tores, l'objectif étant de démontrer le théorème de convexité d'Atiyha-Guillemin-Sternberg. Une discussion d'un théorème de classification de Delzant-Laudenbach est aussi donnée. La présentation se voulant une introduction assez exhaustive à la théorie des actions hamiltoniennes, presque tous les résultats énoncés sont accompagnés de preuves complètes. Divers exemples sont étudiés afin d'aider à bien comprendre les aspects plus subtils qui sont considérés. Plusieurs sujets connexes sont abordés, dont la préquantification géométrique et la réduction de Marsden-Weinstein. / This Master thesis is concerned with some natural notions of group actions on symplectic manifolds, which are in decreasing order of generality : symplectic actions, weakly hamiltonian actions and hamiltonian actions. A knowledge of group actions and of symplectic geometry is a prerequisite ; two chapters are devoted to a coverage of the basics of these subjects. The case of hamiltonian actions is studied in detail in the fourth chapter : the important moment map is introduced and several results on the orbits of the coadjoint representation are proved, such as Kirillov's and Kostant-Souriau's theorems. The last chapter concentrates on hamiltonian actions by tori, the main result being a proof of Atiyah-Guillemin-Sternberg's convexity theorem. A classification theorem by Delzant and Laudenbach is also discussed. The presentation is intended to be a rather exhaustive introduction to the theory of hamiltonian actions, with complete proofs to almost all the results. Many examples help for a better understanding of the most tricky concepts. Several connected topics are mentioned, for instance geometric prequantization and Marsden-Weinstein reduction.

Group actions in Czech procedural law Summary The thesis deals with an issue of group actions, which is a legal instrument of collective protection of private rights in civil proceedings. The aim of this thesis is to analyze this procedural institute, to point out shortcomings of current legislation and to propose it's acceptable solution. The thesis is divided into three relatively independent sections. The first section focuses on theoretical basis (including a brief outline of the historical development) and defines basic terminology used in the thesis. Then it describes the most general division of the collective enforcement mechanisms into a group action and a representative action. The second section concentrates on legislative schemes of group actions in certain foreign jurisdictions. Specifically, it deals with a legal conception of class action in the legal system of the USA, where it is applied as a kind of a opt-out group proceedings (group members, who does not agree with adjudication of their claims, may opt-out), then it deals with opt-in group proceedings in Sweden (group members can be required to enter the suit individually) and finally it describes a German model proceedings in capital market disputes, which represents a compromise between individual and collective proceedings. The third...

L’activité perceptive du sujet semble être impactée par les actions qu’il peut effectivement réaliser à l’égard de son environnement. Néanmoins, il semble que les possibilités d’action du sujet soient le plus souvent envisagées comme des propriétés objectives de l’environnement si bien qu’elles sont généralement décrites comme étant préparatoires à l’action. Ce travail de thèse s’est consacré à défendre l’idée que d’une part, ces possibilités d’action sont des propriétés émergentes de la relation sujet-environnement et que d’autre part, elles sont susceptibles d’avoir un rôle fonctionnel constitutif pour l’activité perceptive. Dans une première série d’études, nous avons observé que des modulations dans les possibilités d’action conduisaient à un renversement des effets de facilitation lorsque des sujets devaient catégoriser des objets préhensibles. Dans une seconde série d’étude, nous avons observé que la capacité à détecter un objet parmi un ensemble d’autres était impactée par les potentiels d’action suggérés par les objets mais également, modulée par l’engagement moteur du sujet dans la tâche. Enfin, dans une troisième série d’études, nous avons mis en évidence que les possibilités d’action pourraient occuper un rôle fonctionnel significatif pour l’activité perceptive du sujet. Ce dernier travail met en effet en évidence que lors de situations perceptives ambiguës, la manière dont est catégorisé un objet semble dépendre de la capacité du sujet à pouvoir le saisir efficacement. L’ensemble de ces résultats semble indiquer que les possibilités d’action sont bien des propriétés du couplage sujet-environnement et occupent une place majeure dans l’activité perceptive. Plus généralement, ces travaux constituent des arguments en faveur d’une prise en compte de l’ensemble de la situation sujet-environnement ainsi que de l’importance de la signification des actions du sujet en fonction des contraintes qui s’exercent sur lui ici et maintenant. / Perceptual activity seems to be impacted by the actions the subject can actually carry out with regard to its environment. Nevertheless, it seems that subject action possibilities are most often considered as objective properties of the environment so that they are generally described as being preparatory to action. This thesis work has been devoted to defending the idea that on the one hand, these possibilities of action are emergent properties of the subject-environment relationship and on the other hand, they are likely to have a constitutive functional role for perceptual activity. In a first series of studies we observed that modulations in the possibilities of action lead to a reversal of the facilitation effects when subjects were to categorize prehensile objects. In a second series of studies we observed that the ability to detect an object among a set of others was impacted by the action potentials suggested by the objects but also modulated by the subject's driving engagement in the task. Finally, in a third series of studies, we have shown that the possibilities of action could play a significant functional role for subject perceptual activity. Indeed, this last work shows that in ambiguous perceptual situations, the way in which an object is categorized seems to depend on the ability of the subject to grasp it effectively. All these results seem to indicate that the action possibilities are properties of the subject-environment coupling and occupy a major place in the perceptual activity. More generally, these works constitute some arguments in favor of taking into account the whole subject-environment situation as well as the importance of the meaning of the subject actions according to the constraints that are exerted on him, here and now.

This study explores how reflection upon oneself and one's own behaviour assists people - and, in particular, managers - to develop. Reflective techniques are examined in the context of action-learning (Revans, 1980 and Marsick, 1992) and are argued to be a powerful means of creating self-understanding, which in turn creates opportunities for self-directed personal change. Reflective techniques are also examined as a means of developing the personal craft or praxis of those who try to assist the development of managers, and as a technique for use in action research (Lewin, 1946) and the development of collective knowledge. Schon's (1987) concept of the 'reflective practitioner' provided a major theoretical foundation for this work. The study employed action research and action learning methodologies. The researcher spent six years honing her understanding and application of reflective techniques in assisting the development of managers. She also applied self-reflection to the development of her own praxis over that time. One result of the study has been the enhancement of the practical, reflection-based techniques used by the writer to facilitate the development of managers - and more importantly, offered to them to facilitate their own continuing development. Hopefully, these techniques will be of value to other practitioners in this field. A second outcome has been the review and refinement of some of the theoretical constructs used by this writer and other practitioners and theorists which help to describe and explain the phenomenon of reflection-based behavioural change. A third outcome has been the documentation of a case-study in the application of reflective techniques to the development of personal praxis, tracking the integration of conceptual understanding and technique. Chapter 1 provides an overview of the rationale, scope, methodology and outcomes of this study. Chapter 2 explores reflection as a technique for research and the development of collective knowledge, and incorporates a review of the relevant literature. Chapters 3 and 5 examine reflection as a tool for learning, drawing on the literature and tracking the development of the researcher's own understanding. Chapter 4 describes how the researcher learned to use reflective learning techniques when working with others and follows the gradual integration of her understanding with her practice. Chapter 6 summarises and reflects upon both the processes and the outcomes of the research.