教学课程资源库 更多>>

关键词： Prediction algorithms   Task analysis   Graph neural networks   Electronic mail   Message passing   Approximation algorithms   Computer science   Graph neural networks (GNNs)   hashing   message passing (MP)   network embedding  

摘要： Given a network, it is well recognized that attributed network embedding represents each node of the network in a low-dimensional space, and, thus, brings considerable benefits for numerous graph mining tasks. In practice, a diverse set of graph tasks can be processed efficiently via the compact representation that preserves content and structure information. The majority of attributed network embedding approaches, especially, the graph neural network (GNN) algorithms, are substantially costly in either time or space due to the expensive learning process, while the randomized hashing technique, locality-sensitive hashing (LSH), which does not need learning, can speedup the embedding process at the expense of losing some accuracy. In this article, we propose the MPSketch model, which bridges the performance gap between the GNN framework and the LSH framework by adopting the LSH technique to pass messages and capture high-order proximity in a larger aggregated information pool from the neighborhood. The extensive experimental results confirm that in node classification and link prediction, the proposed MPSketch algorithm enjoys performance comparable to the state-of-the-art learning-based algorithms and outperforms the existing LSH algorithms, while running faster than the GNN algorithms by 3-4 orders of magnitude. More precisely, MPSketch runs 2121, 1167, and 1155 times faster than GraphSAGE, GraphZoom, and FATNet on average, respectively.

关键词： Computer science   Internet   Artificial intelligence  

摘要： Governing artificial intelligence involves multidisciplinary work across different fields to ensure human control over technology. This article aims to show how computer science can contribute to governance of responsible artificial intelligence technologies and applications.

关键词： Computational modeling   Processor scheduling   Training   Electronic mail   Computer science   Bipartite graph   Analytical models   Multi-server job   online gradient ascent   online scheduling   regret analysis  

摘要： Multi-server jobs that request multiple computing resources and hold onto them during their execution dominate modern computing clusters. When allocating the multi-type resources to several co-located multi-server jobs simultaneously in online settings, it is difficult to make the tradeoff between the parallel computation gain and the internal communication overhead, apart from the resource contention between jobs. To study the computation-communication tradeoff, we model the computation gain as the speedup on the job completion time when it is executed in parallelism on multiple computing instances, and fit it with utilities of different concavities. Meanwhile, we take the dominant communication overhead as the penalty to be subtracted. To achieve a better gain-overhead tradeoff, we formulate an cumulative reward maximization program and design an online algorithm, named OgaSched, to schedule multi-server jobs. OgaSched allocates the multi-type resources to each arrived job in the ascending direction of the reward gradients. It has several parallel sub-procedures to accelerate its computation, which greatly reduces the complexity. We proved that it has a sublinear regret with general concave rewards. We also conduct extensive trace-driven simulations to validate the performance of OgaSched. The results demonstrate that OgaSched outperforms widely used heuristics by 11.33%, 7.75%, 13.89%, and 13.44%, respectively.

关键词： Application programs  

摘要： This paper first presents the NSFC applications and grants of the General Program, Young Scientists Fund and Fund for Less Developed Regions, as well as the Major Program, Excellent Young Scientists Fund, and Distinguished Young Scholars Fund in computer science (F02) in 2023. Then, we analyze the funding situations of universities in computer science in 2023. Finally, the evaluation principles and measures of fund project for computer science are introduced. Meanwhile, the development prospects of computer science in the future are also summarized. © 2024 Science Press. All rights reserved.

关键词： Equity   CS for All  

摘要： Objectives. At the K-12 level, "CS for All" initiatives across the country strive to increase equitable access to and participation in computer science (CS). However, there are many open questions about the implementation and effectiveness of these initiatives, including the extent to which exposing young people to CS early on can shape their longer-term CS interest and engagement. In this article, we examine CS participation among 6th-8th-grade Black girls and assess whether CS participation during middle school shapes CS interest and engagement during their first year of high school. We focus on Black girls in the hopes of developing a more nuanced and rigorous understanding of computing experiences at the intersection of racism and sexism in this field. Participants. The focal group of students in this study are 6th-8th-grade Black girls from New York City. We employ a comparative lens in this article, contextualizing the CS experiences and outcomes of Black girls to Latinas, Asian, and White girls, as well as Black boys. Study Method. We primarily rely on quantitative data for this article, applying a critical lens to our analyses and interpretation. More specifically, we conduct descriptive analyses of course-taking patterns as well as survey data that focus on student attitudes and beliefs about CS. We then carry out inferential analyses of students' administrative records examining how, if at all, middle school CS participation is related to high school outcomes for Black girls. We employ a comparative lens and rely on qualitative data to make sense of our results. Findings. We find that, troublingly, Black girls in the district are disproportionately less likely to receive CS instruction in middle school. Black girls are also less likely than Black boys, Latinas, and White girls to feel that they belong in CS. However, Black girls in CS courses report similar levels of engagement, family, and peer support, as well as value for CS relative to other students i

关键词： Broadening participation in computing   intersectional computing   mentorship  

摘要： Several studies have reported the positive benefits of informal Computer Science learning programs for Black girls, which include staff, mentors, and peers reflective of the girls in the program;however, we do not know enough about what motivates Black women to sign up to teach in such programs, or how representation in mentoring affects future CS learning for Black girls. Here we report the intersectional experiences of 11 Black women who served as instructors and mentors for the virtual INTech Summer Camp. The women in this study confirm that it is imperative to consider the intersectional experiences of Black women in CS as a whole, without explicating their identities. We found that Black women in CS feel that it is their responsibility to give back to Black girls to pay it forward from their own mentorship experiences. The Black women in our study describe seeking to fill a gap in mentorship that they experienced themselves, to share their knowledge and experiences to support the development of others who can relate to that intersection of identifying as Black and a girl or woman. Providing positive representation to encourage young Black girls to pursue tech careers is imperative to diversifying the tech workforce.

关键词： Cryptography   Indexes   Servers   Cloud computing   Privacy   Encryption   Computer science   graph privacy   K-hop reachability query   privacy preserving  

摘要： K-hop Reachability Query (KRQ) is one of fundamental graph queries, which can answer whether a node u can reach a node v within K hops. With the scale of graph data increasing, data owner desires to outsource the local graphs to cloud server. To protect the graph privacy, data owner encrypts graphs before outsourcing them to the cloud server. It imposes a great challenge to KRQ over encrypted graphs. How to realize Privacy-Preserving K-hop Reachability Query (PPKRQ) over encrypted graphs is still an unexplored problem. In this article, we explore this problem and propose a practical scheme. In order to efficiently answer KRQ over encrypted graphs, we construct the encrypted Breadth-First Spanning Tree table and adjacent list D (BFST-D). Based on encrypted BFST table, we can directly judge whether two query nodes are reachable within K hops when they are in one spanning tree. The encrypted adjacent list D can help answer that two query nodes in different spanning trees. To protect the privacy, we utilize the Paillier cryptographic and Order-Revealing Encryption (ORE) to support the comparison and computation over ciphertexts. As a result, our scheme achieves the sensitive information privacy without losing the ability of querying over encrypted graphs. The security analysis shows that our proposed scheme is secure based on semi-honest cloud server. The extensive experiments show the efficiency of our scheme.

关键词： Academic genealogy   Mentorship   Knowledge evolution   Computer science   Knowledge Inheritance   Knowledge mutation  

摘要： Academic genealogy (AG) provides valuable insights into the transmission of knowledge from mentors to mentees, revealing the evolution of knowledge within the academic community. This study explores the intricate dynamics of knowledge evolution within academic genealogies, utilizing on a dataset comprising 16,852 computer science researchers, 613,277 papers, and 11,988 mentorship relationships. By focusing on small-scale knowledge units, our analysis aims to uncover patterns of knowledge inheritance and mutation across different subfields of computer science and highlights several aspects of knowledge evolution in computer science. Firstly, computer science is characterized by strong mentorship ties, indicating the significance of knowledge transmission within the field. Additionally, there is a mix of foundational and developing areas, suggesting a field that is growing and diversifying rather than declining, as indicated by linear regression outcomes. Secondly, our research reveals a surge in collaborative knowledge exchange in computer science since 2000, with fields such as ComputerCommunication Networks and Software Engineering leading in terms of output and impact. Furthermore, areas like Computer Graphics and Artificial Intelligence stand out for their depth and novelty. Thirdly, we categorize researchers into three types: roots, branches, and leaves, reflecting their role in knowledge transmission. Branch researchers tend to innovate, while leaf researchers show a combination of traditional knowledge uptake and new contributions, illustrating the dynamic flow of ideas within the field. Future research endeavors are encouraged to embrace larger datasets and further fortify our understanding of the topic.

关键词： Data models   Computational modeling   Training   Distributed databases   Servers   Data privacy   Computer science   Data elimination   distributed learning   low-quality data   federated unlearning  

摘要： Federated unlearning (FUL) is an emerging distributed machine learning paradigm which enables the removal or unlearning of specific training data effects from trained Federated Learning (FL) models. While current studies mostly focus on client-side FUL to address the "right to be forgotten", and ignore the server's right to remove local models from the global model, particularly when clients are trained with low-quality data. In this paper, we introduce the Server-Initiated Federated Unlearning (SIFU) algorithm, devised to eliminate low-quality data from the global model. SIFU consists of two main components: (i) Identifying low-quality data: we develop a category-based method for quantifying low-quality data for each client and filter out clients containing such data. Datasets are then divided accordingly. (ii) Unlearning low-quality data: we employ gradient ascent training to counteract the adverse effects of low-quality data on local models. To minimize any bias introduced, we concurrently perform several batches of boosting training with good-quality data. SIFU could identify and promptly eliminate the impact of low-quality data on the FL global model while still preserving the benefits of good-quality data. Finally, extensive evaluations are conducted to verify the performance of SIFU with four different kinds of datasets and models. Results show that, compared to retraining from scratch, SIFU accelerates the speed of unlearning by 15x for small datasets (i.e., MNIST and FMNIST) and 20x for large datasets (i.e., CIFAR-10 and CelebA) without any degradation in accuracy, which also outperforms the state of the arts.

关键词： computer science   diversity and inclusion   gender equity   higher education   science education   women in science  

摘要： The gender imbalance in computer science (CS) is one of the most challenging issues in American education. CS is the only science, technology, engineering, and mathematics (STEM) field in which women's representation has steadily declined in recent decades. In this study, we explored one potential approach that could be effective in increasing college women's participation in CS: participation in Women-in-Computing (WiC) groups. Through participant observation and individual interviews in a WiC group at a major research university, we investigated how students engage in WiC, the impacts of the WiC on identity and belonging, and the challenge of sustainability. The results were coded using a hybrid of grounded and deductive coding and indicate that WiC groups offer various programs and events that enable women in CS to fully participate, learn, and grow. WiC represents an identity, a community, a safe space, and a journey. The results also suggest that the WiC has had positive impacts on students' identity and belonging, as evidenced by increased self-efficacy, reduced imposter syndrome, and enhanced sense of belonging and community. Furthermore, we outline three strategies employed by the WiC to ensure the group's sustainability. Our study sheds light on how WiC can encourage women to enter and persist in CS, and on some of the characteristics of a successful WiC. We demonstrate that WiC may be potentially effective in diversifying CS through identity-based participation. Moreover, student leaders design both the structure of the group and the leadership continuity process to ensure sustainability.