State-of-the-art temporal action detectors inefficiently search the entire video for specific actions. Despite the encouraging progress these methods achieve, it is crucial to design automated approaches that only explore parts of the video which are the most relevant to the actions being searched. To address this need, we propose the new problem of action spotting in videos, which we define as finding a specific action in a video while observing a small portion of that video. Inspired by the observation that humans are extremely efficient and accurate in spotting and finding action instances in a video, we propose Action Search, a novel Recurrent Neural Network approach that mimics the way humans spot actions. Moreover, to address the absence of data recording the behavior of human annotators, we put forward the Human Searches dataset, which compiles the search sequences employed by human annotators spotting actions in the AVA and THUMOS14 datasets. We consider temporal action localization as an application of the action spotting problem. Experiments on the THUMOS14 dataset reveal that our model is not only able to explore the video efficiently (observing on average 17.3% of the video) but it also accurately finds human activities with 30.8% mAP (0.5 tIoU), outperforming state-of-the-art methods

The doctoral thesis Looking for the Sense of Being as a Way of Daseinsanalytic Psychotherapy with the subtitle Potential Contribution of Daseinsanalysis for Pedagogical Practice deals with possibilities of the daseinsanalytical preventive and treatment consultancy application in the environment of primary schools. The consultation programmes are based on Martin Heidegger's philosophy, phenomenology, hermeneutical explanation method and daseinsanalytical psychotherapy. The goal of the doctoral thesis is to introduce, in the field of philosophy of education, this new project focused on bullying and also to raise a question about the inspirational possibilities of the development of the current system of Teacher's university education. KEYWORDS Philosophy of education, daseinsanalysis, bulliyng, prevention, consulting, education in teaching

The artist discusses the work in Silly Trip Wires, 2020 his Master of Fine Arts exhibition. The exhibition includes an installation, Silly Trips Wires, and documentation of a smaller site-specific version of the work. The Artist discusses the process of transition from military to civilian, and the potential effects that mental trauma from combat deployments can have on this process. This is tied to an analysis of how communicating the experience of veterans to civilians, through artwork, functions to bring about understanding.

Just like adults, children of all ages need time and understanding in order to process the concept of death and dying. This process is much different for children than it is for adults. There are 4 components relative to children's understanding of death: (a) the irreversibility factor, (b) finality, (c) inevitability, and (d) causality. These 4 components relate directly to the developmental level of the child at the time the death occurs. Knowing how children's concept of death is constructed provides parents and caregivers important information and helps them respond more sensitively to what children might feel and experience. This article provides an overview of how children under-stand death, concrete strategies for talking to children about death, and suggestions for teachers about how to help children through grief and mourning.

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第21438号 / 情博第689号 / 京都大学大学院情報学研究科社会情報学専攻 / (主査)教授 吉川 正俊, 教授 神田 崇行, 教授 森 信介 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM

This research study investigated how professional mathematicians understand and operate with highly-abstract, advanced mathematical concepts in their own work. In particular, this study examined how professional mathematicians operated with mathematical concepts at different levels of understanding. Moreover, this study aimed to capture what factors influence professional mathematicians' level of understanding for particular mathematical concepts. To frame these research goals, three theoretical levels of understanding were proposed, process-level, pseudo-object-level, object-level, leveraging two ways that Piaget (1964) described what it meant to know or understand a mathematical concept. Specifically, he described understanding an object as being able to "act on it," and also as being able to "understand the process of this transformation" (p. 176). Process-level understanding corresponds to only understanding the underlying processes of the concept. Pseudo-object-level understanding corresponds to only being able to act on the concept as a form of object. Object-level understanding corresponds to when an individual has both of these types of understanding. This study is most especially concerned with how professional mathematicians operate with a pseudo-object-level understanding, which is called pseudo-objectification. For this study, six professional mathematicians with research specializing in a subfield of algebra were each interviewed three times. During the first interview, the participants were given two mathematical tasks, utilizing concepts in category theory which were unfamiliar to the participants, to investigate how they operate with mathematical concepts. The second interview utilized specific journal publications from each participant to generate discussion about influences on their level of understanding for the concepts in that journal article. The third interview utilized stimulated recall to triangulate and support the findings from the first two interviews. The findings and analysis revealed that professional mathematicians do engage in pseudo-objectification with mathematical concepts. This demonstrates that pseudo-objectification can be productively leveraged by professional mathematicians. Moreover, depending on their level of understanding for a given concept, they may operate differently with the concept. For example, when participants utilized pseudo-objects, they tended to rely on figurative material, such as commutative diagrams, to operate on the concepts. Regarding influences on understanding, various factors were shown to influence professional mathematicians' level of understanding for the concepts they use in their own work. These included factors pertaining to the mathematical concept itself, as well as other sociocultural or personal factors. / Doctor of Philosophy / In this research study, I investigated how professional mathematicians utilize advanced mathematical concepts in their own work. Specifically, I examined how professional mathematicians utilize mathematical concepts that they do not fully understand. I also examined what factors might influence a professional mathematician to fully understand or choose not to fully understand a mathematical concept they are using. To address these goals, six research-active mathematicians were each interviewed three times. In these interviews, the mathematicians engaged with mathematical concepts that were unfamiliar to them, as well as concepts from one of their own personal research journal publications. The findings demonstrated that professional mathematicians sometimes utilize mathematical concepts in different ways depending on how well they understand the concepts. Moreover, even if mathematicians do not have a full understanding of the concepts they are using, they can still sometimes productively leverage this amount of understanding to successfully reach their goals. I also demonstrate that various factors can and do influence how well a professional mathematician understands a given mathematical concept. Such influences could include the purpose of use for the concept, or what a mathematician's research community values.

The automatic analysis of emotions is a well-established area in the natural language processing ( NLP ) research field. It has shown valuable and relevant applications in a wide array of domains such as health and well-being, empathetic conversational agents, author profiling, consumer analysis, and security. Most emotion analysis research till now has focused on sources such as news documents and product reviews. In these cases, the NLP task is the classification into predefined closed-set emotion categories (e.g. happy, sad), or alternatively labels (positive, negative). A deep and fine-grained emotion analysis would require explanations of the trigger events that may have led to a user state. This type of analysis is still in its infancy. In this work, we introduce the concept of Emotion Carriers (EC) as the speech or text segments that may include persons, objects, events, or actions that manifest and explain the emotions felt by the narrator during the recollection. In order to investigate this emotion concept, we analyze Personal Narratives (PN) - recollection of events, facts, or thoughts from one’s own experience, - which are rich in emotional information and are less explored in emotion analysis research. PNs are widely used in psychotherapy and thus also in mental well-being applications. The use of PNs in psychotherapy is rooted in the association between mood and recollection of episodic memories. We find that ECs capture implicit emotion information through entities and events whereas the valence prediction relies on explicit emotion words such as happy, cried, and angry. The cues for identifying the ECs and their valence are different and complementary. We propose fine-grained emotion analysis using valence and ECs. We collect and annotate spoken and written PNs, propose text-based and speech-based annotation schemes for valence and EC from PNs, conduct annotation experiments, and train systems for the automatic identification of ECs and their valence.

<p>This thesis examines the societal forces which have helped shape the present-day form of the institutionalization, criticism, and appreciation of art. Specifically, it examines the influence of modern thought on our present understanding of art. First, we examine how modernists have typically 100ked at art. Both the enlightening aspects as well as the deficiencies of modernist aesthetics are uncovered. Also, with the help of Jurgen Habermas, we examine a modernist societal approach to aesthetics. Second, the fundamental philosophical presuppositions of modernity are uncovered so that the societal forces that have helped to make art an autonomous institutionalized field of expertise can be examined. In passing, we discuss the concept of "lifeworld". We then examine the explanatory powers of considering the arts as forms of language. Third, as Habermas's social theory indicates, an excursion into the theory of argumentation provides indications of the mechanisms involved in the understanding of art. We consider the rhetorical, dialectical, and logical aspects of both non-verbal and linguistic argumentation. This provides us with a forum for discussing Habermas' s notion of an ideal speech situation, Gadamer's concept of the various modes of iii experiencing tradition and its parallel with the experiencing of art, and Quine's thesis of the indeterminacy of translation and what it implies for non-verbal forms of art. Fourth, we examine the implications and explanatory powers of Habermas's three-world distinction, which is, in turn, derived from the modernist presupposition of the distinction between subject and object. With these distinctions, we can see that the existence of a highly specialized field of expertise surrounding art and notions such as "art for art's sake" are not accidental. To conclude, we examine emotional life and its implications for modern notions of art.</p> / Master of Arts (MA)

Atmospheric Flight Mechanics (AFM) is one of the cornerstones of aeronautical engineering and includes subjects like aerodynamic prediction, stability and control, dynamics, and vehicle design. These topics are critical to the success of aircraft development, so AFM is considered one of the most important foundational knowledge areas for aerospace engineering. Unfortunately, students graduating from aerospace engineering programs are often underprepared to perform in AFM jobs. This ongoing research focuses on developing a blueprint for assessing conceptual understanding of AFM concepts. Since existing literature suggests that novices and experts organize knowledge differently, comparing students' and experts' mental models can shine a light on the alternative conceptions that students retain post-instruction. As such, framing the study around synthetic mental models can be advantageous. To explore these mental models, three types of data have been collected and analyzed. Document analysis was done on course documents to identify what concept relationships were being presented to the students. Class observations were conducted to analyze how concepts are introduced to students and what relationships are highlighted by the instructor. Finally, a concept mapping activity was facilitated to study the mental models that the students built after instruction. The results show a lack of synthetization between the knowledge introduced in the classroom and students' prior knowledge which translated into student mental models that do not meet some of the expectations of the course. Moreover, this study highlights the importance of the instructor's awareness of their own expectations for learning and knowledge synthetization in the design of an AFM course. / Doctor of Philosophy / Conceptual understanding research has often focused on how students develop their understanding of scientific concepts that are difficult to grasp. Through this research, many assessment techniques have been developed and implemented in the design of STEM courses. However, many of these techniques and implementations have been limited to K-12 or introductory engineering courses. Atmospheric Flight Mechanics (AFM) is an important part of the aerospace curriculum that has yet to be studied under the conceptual understanding lens. The goal of this study was to investigate how students develop AFM conceptual understanding using a synthetic mental model framework. This study focused on answering three questions, how are students being introduced to AFM concepts?, how do students' mental models develop throughout the semester?, and how do the students' and instructor's mental models compare?. Through the exploration of class documents, class observations, and concept mapping activities, this research found that students are having a hard time making sense of new knowledge based on their previous understanding of similar topics. By trying to integrate this new knowledge with their previous mental models, students are developing synthetic mental models that do not align with the scientific explanations of the topic. This study also found that instructors are often unaware of their own knowledge and expectations for learning which makes knowledge synthetization harder for the students. Therefore, addressing these issues during course design could make an AFM course easier to understand for students.

With the rapid development of computer and information technologies, medical imaging has become one of the major sources of information for therapy and research in medicine, biology and other fields. Along with the advancement of medical imaging techniques, computer-aided detection and diagnosis (CAD/CADx) has recently emerged to become one of the major research subjects within the area of diagnostic radiology and medical image analysis. This thesis presents two multi-level learning-based approaches for medical image understanding with applications of CAD/CADx. The so-called "multi-level learning strategy" relies on that supervised and unsupervised statistical learning techniques are utilized to hierarchically model and analyze the medical image content in a "bottom up" way. As the first approach, a learning-based algorithm for automatic medical image classification based on sparse aggregation of learned local appearance cues is proposed. The algorithm starts with a number of landmark detectors to collect local appearance cues throughout the image, which are subsequently verified by a group of learned sparse spatial configuration models. In most cases, a decision could already be made at this stage by simply aggregating the verified detections. For the remaining cases, an additional global appearance filtering step is employed to provide complementary information to make the final decision. This approach is evaluated on a large-scale chest radiograph view identification task and a multi-class radiograph annotation task, demonstrating its improved performance in comparison with other state-of-the-art algorithms. It also achieves high accuracy and robustness against images with severe diseases, imaging artifacts, occlusion, or missing data. As the second approach, a learning-based approach for automatic segmentation of ill-defined and spiculated mammographic masses is presented. The algorithm starts with statistical modeling of exemplar-based image patches. Then, the segmentation problem is regarded as a pixel-wise labeling problem on the produced mass class-conditional probability image, where mass candidates and clutters are extracted. A multi-scale steerable ridge detection algorithm is further employed to detect spiculations. Finally, a graph-cuts technique is employed to unify all outputs from previous steps to generate the final segmentation mask. The proposed method specifically tackles the challenge of inclusion of mass margin and associated extension for segmentation, which is considered to be a very difficult task for many conventional methods. / Master of Science