教学课程资源库 更多>>

关键词： Federated learning   Feature extraction   Big Data   Collaboration   Graph convolutional networks   Computer science   Accuracy   Feature fusion   graph convolutional networks   online distillation   sequential recommendation  

摘要： Sequential recommendation seeks to understand user preferences based on their past actions and predict future interactions with items. Recently, several techniques for sequential recommendation have emerged, primarily leveraging graph convolutional networks (GCNs) for their ability to model relationships effectively. However, real-world scenarios often involve sparse interactions, where early and recent short-term preferences play distinct roles in the recommendation process. Consequently, vanilla GCNs struggle to effectively capture the explicit correlations between these early and recent short-term preferences. To address these challenges, we introduce a novel approach termed Graph Convolutional Networks with Collaborative Feature Fusion (COFF). Specifically, our method addresses the issue by initially dividing each user interaction sequence into two segments. We then construct two separate graphs for these segments, aiming to capture the user's early and recent short-term preferences independently. To obtain robust prediction, we employ multiple GCNs in a collaborative distillation manner, incorporating a feature fusion module to establish connections between the early and recent short-term preferences. This approach enables a more precise representation of user preferences. Experimental evaluations conducted on five popular sequential recommendation datasets demonstrate that our COFF model outperforms recent state-of-the-art methods in terms of recommendation accuracy.

关键词： Computer science   Engineering  

摘要： Completing complex tasks in unpredictable settings challenges robotic systems, requiring a step change in machine intelligence. Sensorimotor abilities are considered integral to human intelligence. Thus, biologically inspired machine intelligence might usefully combine artificial intelligence with robotic sensorimotor capabilities. Here we report an embodied large-language-model-enabled robot (ELLMER) framework, utilizing GPT-4 and a retrieval-augmented generation infrastructure, to enable robots to complete long-horizon tasks in unpredictable settings. The method extracts contextually relevant examples from a knowledge base, producing action plans that incorporate force and visual feedback and enabling adaptation to changing conditions. We tested ELLMER on a robot tasked with coffee making and plate decoration;these tasks consist of a sequence of sub-tasks from drawer opening to pouring, each benefiting from distinct feedback types and methods. We show that the ELLMER framework allows the robot to complete the tasks. This demonstration marks progress towards scalable, efficient and 'intelligent robots' able to complete complex tasks in uncertain environments.

关键词： Resource management   Spread spectrum communication   Multi-agent systems   Robots   Robot kinematics   Time factors   Collaboration   Computer science   Optimization   Deep reinforcement learning   Task allocation   deep reinforcement learning   task exploration   network connectivity  

摘要： For a limited scale self-organized multi-agent system operating in environments with unknown task distributions, one challenge is to reduce the task response time via efficiently combining task exploration and allocation, another challenge is to improve the task completion rate via unlocking the potential of network cooperation in task allocation. However, in the existing studies, task allocation is generally regarded as an independent issue for known task distribution environments, rarely combined with task exploration, also hardly solving the conflict between the multi-hop network cooperation and mobility flexibility of agents. In view of this, this paper proposes a network connectivity-aware deep reinforcement learning method for task exploration and allocation in limited scale multi-agent systems (NCADRL4TEA). This method divides the task environment into regions and integrates task exploration with task allocation via two policies: a leaving policy to guide global task exploration among regions according to the distribution of agents and tasks, and a stay policy to guide local task allocation within each region according to the multi-hop network cooperation performance between agents. Further, in the stay policy, a network connectivity-aware task allocation optimization model is provided, which leads agents in the same region to cooperate with each other via multi-hop intermittent network connectivity and flexibly adjust their locations until the optimal multi-hop network cooperation performance is achieved. The experimental results verify that NCADRL4TEA can reduce the task response time in combination of task exploration and allocation, and improve the task completion rate in network cooperation.

关键词： Crowdsourcing   Social networking (online)   Mobile computing   Recruitment   Computer science   Spatiotemporal phenomena   Predictive models   Faces   Sensors   Prediction algorithms   Mobile crowdsourcing   cold start problem   social networks   user retention  

摘要： Mobile crowdsourcing (MCS) has attracted widespread attention by recruiting users with mobile devices to collect crowdsourcing data. Existing research on MCS assumes that the platform has sufficient users. However, platforms in their early stages of development face the cold start problem, which can lead to their inability to grow or even result in bankruptcy. While some studies try to solve it by recruiting users through social networks to participate in crowdsourcing tasks, they only focus on how to recruit more users without addressing the issue of user retention. This can lead to an increasing proportion of users losing interest in the platform and dropping out and thus it fails to solve the cold start problem truly. In light of this, we present a task recommendation-based method to recruit new users via the social network and keep registered users active on the platform. Specifically, we first use an extended independent cascade model to describe the recruitment of users through social networks. Secondly, we use a task acceptance model to describe user decisions. Finally, we utilize a fuzzy control system that incorporates spatiotemporal crowdsourcing information to predict user behaviour and recommend tasks to users most likely to complete them. Extensive experiments on large-scale real datasets were conducted to evaluate the proposed solution. The results indicate that compared to existing methods such as SocialRecruiter, our solution reduces the 30-day average user churn rate by 23.90% while significantly boosting user retention and task completion rates by up to 23.73% and 48.7%, respectively.

关键词： Cognitive diagnostic modeling   Programming education   Information technology   Computer science   Educational assessment  

摘要： The global shortage of skilled programmers remains a persistent challenge. High dropout rates in introductory programming courses pose a significant obstacle to graduation. Previous studies highlighted learning difficulties in programming students, but their specific weaknesses remained unclear. This gap exists due to the predominant focus on the overall academic performance evaluation. To address this gap, this study employed cognitive diagnostic modeling (CDM) to profile the skill mastery of programming students. An empirical analysis was conducted to select the most appropriate model for the data, and the linear logistic model (LLM) was determined to be the best fit. Final examination results from 308 information technology (IT) and 279 computer science (CS) students were analyzed using the LLM. Unfortunately, findings revealed that programming students exhibited proficiency primarily in code tracing and language proficiency but displayed deficits in theoretical understanding, logical reasoning, and algorithmic thinking. From a practical standpoint, this deficiency in fundamental skills sheds light on the factors contributing to academic failures and potentially eventual dropout in programming education. When comparing the student population by academic program, CS students demonstrated superior mastery compared to their IT counterparts, although both groups exhibited a lack of mastery in code tracing. These deviations underscore the pressing need for tailored educational strategies that address the unique strengths and weaknesses of each student group. Overall, this study offers valuable insights into programming education literature and contributes to the expanding application of CDM in educational research.

关键词： Quantum entanglement   Routing   Quantum computing   Photonics   Optical switches   Repeaters   Throughput   Qubit   Optimization   Computer science   Quantum networks   entanglement routing   EPS assignment   all-optical switching  

摘要： Entanglement routing plays a vital role in distributed quantum computing and quantum networks. Previous works on entanglement routing have addressed several design challenges due to limited quantum resources, failures to establish entanglement, and decoherence of established entanglement, but considered neither the limitations imposed by having a limited number of entangled photon sources (EPSes), nor the benefits of using all-optical (or quantum) switching to distribute entangled photons. In this paper, we first explore the problem of jointly optimizing entanglement routing and EPS assignment, assuming no all-optical switching capability. In other words, a pair of entangled photons generated by one EPS can be distributed to two neighboring quantum nodes. We then relax the above assumption so as to allow a pair of entangled photons generated by one EPS to be distributed to non-adjacent nodes using all-optical switching. We propose two corresponding solutions, namely, entanglement routing and EPS assignment (or ERSA), and ERSAwith all-optical switching-based distribution (or ERSA $+$ D) that aim to maximize the number of entanglement connections between the given set of source-destination (SD) pairs while avoiding starvation and achieving fairness among the SD pairs. In order to obtain efficient solutions in a large discrete solution space in a timely manner, we first formulate each optimization problem as an Integer Linear Programming (ILP), and then propose efficient algorithms to derive near-optimal solutions based on relaxation, Lagrangian decomposition, duality iteration, and rounding techniques. Extensive simulations show that ERSA $+$ Dcan increase network throughput by up to 1652% and 113%, respectively, thanks to EPS assignment optimization, and all-optical switching based entanglement distribution.

关键词： talent development   enrichment   mentoring   social/emotional   computer science  

摘要： Technology is an increasingly omnipresent and rapidly changing component of today's world. As a result, there is an urgency to develop talent in the field of computer science (CS). The purpose of this article is to examine ways that educational systems can develop students' talents to fill the need for skilled CS professionals. We provide an overview of the talent development megamodel as an exemplar of talent development frameworks and discuss ways to tailor it to fit CS talent trajectories. We describe three key components in a CS talent trajectory: early access to materials and programs;support from mentors;and the development of psychosocial skills. We introduce a proposed model, the computer science talent development model (CSTDM), for finding and developing talents in computer science in K-12 settings. CSTDM includes advanced STEM learning opportunities beginning in middle school, a focus on the development of psychosocial skills, and guidance for students regarding STEM career development.

摘要： STEM GIrls (STEM = Science, Technology, Engineering, Mathematics; GIrls = Wortspiel aus Gesellschaft für Informatik (GI) und Mädchen/Frauen als Hauptzielgruppe) ist eine ehrenamtlich organisierte, digitale Workshopreihe der Jungen Gesellschaft für Informatik, die durch weibliche Vorbilder, Vernetzung und geschützten Austausch das Ziel verfolgt, mehr Frauen und nichtbinäre Personen zu MINT-Berufen zu ermutigen. Im Vordergrund stehen dabei berufliche Möglichkeiten in der Informatik. Mit 340 Anmeldungen, 83,7 % Zufriedenheit und einem Net Promoter Score von 55,81 war die Durchführung im Jahr 2023 ein voller Erfolg. Dabei wurden vorwiegend neue Interessierte angesprochen, aber auch einige Personen, die bereits im Vorjahr teilgenommen hatten. Fast 40 % der Teilnehmenden wurden durch STEM GIrls zu einer Mitgliedschaft in der Gesellschaft für Informatik ermutigt oder zumindest dazu angeregt, über eine GI-Mitgliedschaft nachzudenken. Gleichzeitig wurden bestehende GI-Mitglieder motiviert, GI-Mitglied zu bleiben und weiterhin eine Karriere in der Informatik zu verfolgen. Die Auswertung aus 2023 ergab außerdem, dass STEM GIrls die berufliche Orientierungslosigkeit der Teilnehmenden signifikant mindert. Deskriptiv konnte nach STEM GIrls 2023 auch ein Anstieg des Interesses an MINT-Berufen sowie der beruflichen Selbstwirksamkeit bei den Teilnehmenden beobachtet werden. Somit leistet STEM GIrls einen wichtigen Beitrag für mehr Geschlechtervielfalt innerhalb der Gesellschaft für Informatik sowie der Informatik im Allgemeinen.

关键词： Face recognition   Faces   Convolution   Image restoration   Generative adversarial networks   Accuracy   COVID-19   Semantics   Robustness   Computer science   Face de-occlusion   masked face dataset   face recognition   structure-guided   generative adversarial network  

摘要： During the COVID-19 coronavirus epidemic, almost everyone wears a facial mask, which poses a huge challenge for face recognition. Therefore, it is urgent to improve the performance of face de-occlusion of masked faces. However, previous inpainting approaches are limited as they require the knowledge of a given mask, and in the past there was also a lack of real-world masked face datasets suitable for the face de-occlusion task. To tackle above issues, we pioneer a real-world masked face de-occlusion dataset (RMFDD) with accurate mask labels. Further, we propose an occlusion aware and structure-guided network (OASG-Net) for face de-occlusion, consisting of mask prediction subnet, structure prediction subnet, and face de-occlusion subnet. In particular, due to the mask prediction subnet, OASG-Net achieves face de-occlusion without a given external mask. We also use face structure to guide OASG-Net, which makes the recovered face topology more natural and realistic. Besides, we design the mask aware layer to avoid the hard 0-1 mask updating of partial convolution in the face de-occlusion subnet. Extensive results on both face de-occlusion and face recognition tasks demonstrate the superiority of our OASG-Net over the state-of-the-art competitors. Code is available at https://***/WHUfreeway/OASG-Net-Occlusion-Aware-and-Structure-Guided-Network-for-Face-De-Occlusion.

关键词： Computer science   Digital badges   Computational thinking (CT) skills   CT self-efficacy   Learning motivation   Primary education  

摘要： Computational thinking (CT) is regarded as the conceptual basis required for identifying and solving real-world problems through algorithmic approaches to achieve transferable solutions. Developing students' CT in primary school stages is essential and has received considerable attention. Meanwhile, researchers incorporated digital badges into classrooms to improve students' learning motivation in computer science (CS). However, few solid empirical studies analyze the effects of the digital badges approach in CS education to develop students' CT. Therefore, to bridge this gap, this study designed and implemented a block-based programming course in a primary school and examined the impact of the digital badges approach on pupils' CT skills and self-efficacy and learning motivation in CS. A quasi-experimental design was conducted among 99 third-grade students across three classes. The results revealed that the adapted digital badges in primary school (1) significantly improved the students' CT skills, (2) had no significant positive impact on their CT self-efficacy or learning motivation in CS, and (3) showed no mediating effect of learning motivation on the relationship between CT skills and CT selfefficacy. In addition, we found that whether students could see only their own digital badges or also those of others had no impact on CT skills, self-efficacy, or learning motivation in CS. These findings have implications for the design of digital badge educational interventions intended to promote students' CT and learning motivation.