教学课程资源库 更多>>

摘要： Die Erfindung des universellen Komputers hat für die Naturwissenschaft insgesamt eine völlig neue Welt eröffnet, in der eine neue Naturwissenschaft entstanden ist, die vor allem von der Künstlichen Intelligenz (KI) als Disziplin repräsentiert wird. In dieser Arbeit wird sie inhaltlich genauer als Theoriebildung über repräsentierende Objekte (ROBs) charakterisiert, welche sich umfassend nur mit Komputern experimentell überprüfen läßt. Dies wird an unterschiedlichsten Beispielen illustriert, herausragende Aspekte ihrer historischen Entwicklung in den letzten hundert Jahren werden aufgezeigt und ihr aktueller Status wird problematisiert.

关键词： Computer science   Ethics   Education   Machine learning   Cognition   Information and communication technology   Cultural differences  

摘要： There is a dichotomy in computing education. The information communication technology (ICT) field, which includes all areas that pertain to technology, computing, or computational reasoning (e.g., computer science, computer engineering, and machine learning), needs diversity to thrive. Yet, the undergraduate programs that support the field have a difficult time attracting and retaining that diversity [1]. In undergraduate education, ICT is one of the most exclusionary of the science, technology, engineering, and mathematics (STEM) cultures [2] for women. Exclusion challenges a woman's sense of belonging [3] and identity-her sense of "personal relevance, ownership, and integration into the sense of self " [4, p. 208]. Significantly, women's intersectional identities (e.g., the intersection of race, gender, and disability) frame their experiences with oppression, causing problematic experiences in ICT education to compound and hurt computing identity [5].

关键词： Proposals   Fuses   Location awareness   Reliability   Computational modeling   Image segmentation   Computer science   Appearance fusion   Gaussian mixture model   saliency   co-saliency   co-localization  

摘要： Various types of saliency cues exist, all of which can be instrumental in the foreground extraction. It brings us to an interesting problem of effectively combining them. Note that earlier works either fuse them in the spatial domain or introduce dedicated terms in the energy functions to cater to multiple cues. In contrast, this paper investigates the appearance domain and proposes a novel appearance fusion framework, which we refer to as AppFuse. It is an intuitive framework for fusing candidate appearance models into the desired one for an energy function. Thus, we do not require any alterations in the energy function anymore. Like any fusion strategy, the proposed framework also requires guidance, which we facilitate through reliability and mutual consensus phenomena. To demonstrate the efficacy, we leverage it to solve a foreground extraction problem named video co-localization, where we propose two novel concepts i) hierarchical co-saliency and ii) mask-specific proposals. Our fusion results ensure that similar objects get highlighted sufficiently to ensure localization simply by respecting our framework and different spatiotemporal constraints. Our exhaustive set of experiments using both hand-crafted and learned saliency cues reveal that our approach comfortably outperforms several competing localization methods on standard benchmark datasets.

关键词： Ethics   pedagogy   computer science   social justice  

摘要： We studied the impact of introducing first-year computer science (CS) students to ethical thinking about the social justice impacts of data collection, tracking, bias, internet privacy, and competitive "real world" system design and critique activities. While basic content was consistent for all, one group was involved throughout the course in peer discussions designed to foster greater engagement, with the anticipation that this would enable students to reach new levels of sensitivity through peer-to-peer interaction. This article reports on our observation of this design, interview, and project data collected throughout the course as well interviews conducted eight months later to learn about how students were retaining and applying what they learned. We found that students are sensitive to the technology-related risks and vulnerabilities encountered by individuals based on race, gender, and, to some extent, age, but they struggle to assess who is responsible for these risks, what to do about bias in technology design, and how to mitigate harms for individuals whom they perceive to be vulnerable, furthering the argument for an integrated ethics curriculum. We explore the value of formal peer-led discussion to evolve social justice thinking with a focus on identity, though note that opportunities for any group discussion are meaningful to students' thinking about social justice. Over the longer term, students tend to recall and apply ethics that is closely related to their identity, suggesting that empathy has limits.

关键词： Ecosystems  

摘要： With the transition of innovation to the digital sphere, software has become an essential part of contemporary inventions and creations. It is also an essential part of intellectual property. The state of software protection in the European innovation ecosystem still needs to be considered fully regulated. However, researchers in computer science also face several challenges when exploiting intellectual property rights in the software. The field, therefore, offers opportunities for researching, which should steer research to (1) present the critical points, and (2) update some of the legal frameworks to address more clearly the field that deals with the issue of remuneration for researchers in computer science. The overview paper discusses software, focusing mainly on the challenges that researchers in computer science face in protecting and licensing software in the European innovation ecosystem. The following paper presents the state of the researchers' remuneration under protection and licensing. © 2022 Authors. All rights reserved.

关键词： biological design argument   conjunctive explanation   contingency   design   evolution   theistic evolution  

摘要： Rope Kojonen has presented a novel argument for design in biology by drawing on insights from evolutionary science. Without objecting to the explanatory role of evolution, he argues that there is further explanatory work to be done and that this is best achieved by an appeal to design. Here, I interpret his argument, and attempt to evaluate it, as a conjunctive explanation since he appeals to two explanations to account for the purposeful order and complexity of living organisms. Understood in this way, I argue that he makes a very plausible case that two explanations (evolution and design) are better than one (evolution). I also explore the role that direct guidance by God might play in such an argument.

关键词： Computer science student   Dropping out   Factor analysis   Intervention   Studying method  

摘要： Currently, the dropout rate is crucial in the field of Computer Science (CS) higher education. In CS education it is usually the mathematically oriented subjects that are blamed for the high dropout rates. Implementing a theoretical framework into practice, we have been able to prevent 28% of our students from dropping out in the last 2 years due to our education reform. The aim of the present study is to analyse the results of the students through the curriculum of the CS program by factor analysis. Nearly 4000 first-year students’ results were analysed. One of the most important steps of the education reform was that all of the lectures became compulsory to attend. Another step was the introduction of a prevention and skills-training program for every first-year student in order to develop their study skills. Our findings highlighted that as a consequence of the education reform, more students stayed until the end of the first and second semesters and try taking exams in the exam period. Analysing the subjects as factors in the CS curriculum could (1) help faculty staff introduce an education reform, and (2) help decision-makers develop prevention and promotion programs in order to develop students’ study skills. The results reveal that we have managed to successfully engage first year students in the academic environments. © 2022, The Author(s).

关键词： Computing education   education climate   inclusion   diversity   microaggression   scenarios   vignettes  

摘要： Helping students learn to identify and respond to situations involving discrimination is important, especially in fields like Computer Science where there is evidence of an unwelcoming climate that disproportionately drives underrepresented students out of the field. While students should not be considered responsible for fixing issues around discrimination in their institutions, they do have a role to play. In this paper, we present the results of a study in which 318 undergraduate computer science majors were presented with scenarios of discrimination and asked to identify the issues, rate the severity of the issues, and ideate 3-5 responses to address the described situations. They were also asked to identify which of their responses would likely be most effective in addressing discrimination and which of their responses they would be most likely to use if they were in the situation described in real life. Our results show that while students generally are able to identify various forms of discrimination (sexism, racism, religious discrimination, ethnic discrimination, etc.), any ambiguity in a scenario led to students describing the scenario as less severe and/or as an example of oversensitivity. We also show that students come up with many passive responses to scenarios of discrimination (such as ignoring the situation or wishing it had not happened in the first place). Students in our study were more likely to say they would deploy passive responses in real life, shying away from responses that involve direct confrontation. We observed some differences between student demographic subgroups. Women and BIPOC students in CS tend to think these issues are more severe than men andWhite and Asian students in CS. Women are more likely to ideate direct confrontation responses and report willingness to use direct confrontation responses in real situations. Our work contributes a methodology for examining student awareness and understanding of diversity issues as well a

关键词： argument from design   Charles Darwin   evolution   Charles Lyell   natural theology   proximate-ultimate distinction   scientific explanation  

摘要： This article engages Kojonen's discussion of scientific explanation. Kojonen claims the best way to conceptualize the relationship between evolutionary explanations and explanation by design is through the proximate-ultimate distinction and the levels metaphor. However, these are not robust explanatory models but examples of how one might differentiate ambiguous explananda contained in why-questions. Disambiguating explananda is a helpful tool for determining when a situation calls for further explanation;however, on this picture, that some further explanation is needed does not, as proponents of design arguments assume, specify design. The question of whether design is a good explanation at all hinges on what precisely we want explained.

关键词： Training   Generators   Adaptation models   Task analysis   Computer science   Computational modeling   Probabilistic logic   Adversarial learning   graph neural network   inductive learning   negative sampling (NS)   network embedding  

摘要： Network representation learning (NRL) has far-reaching effects on data mining research, showing its importance in many real-world applications. NRL, also known as network embedding, aims at preserving graph structures in a low-dimensional space. These learned representations can be used for subsequent machine learning tasks, such as vertex classification, link prediction, and data visualization. Recently, graph convolutional network (GCN)-based models, e.g., GraphSAGE, have drawn a lot of attention for their success in inductive NRL. When conducting unsupervised learning on large-scale graphs, some of these models employ negative sampling (NS) for optimization, which encourages a target vertex to be close to its neighbors while being far from its negative samples. However, NS draws negative vertices through a random pattern or based on the degrees of vertices. Thus, the generated samples could be either highly relevant or completely unrelated to the target vertex. Moreover, as the training goes, the gradient of NS objective calculated with the inner product of the unrelated negative samples and the target vertex may become zero, which will lead to learning inferior representations. To address these problems, we propose an adversarial training method tailored for unsupervised inductive NRL on large networks. For efficiently keeping track of high-quality negative samples, we design a caching scheme with sampling and updating strategies that has a wide exploration of vertex proximity while considering training costs. Besides, the proposed method is adaptive to various existing GCN-based models without significantly complicating their optimization process. Extensive experiments show that our proposed method can achieve better performance compared with the state-of-the-art models.