教学课程资源库 更多>>

关键词： Quality assessment   Accuracy   Minimization   Mathematical models   Videos   Training   Target recognition   Computer science   Computational modeling   Artificial intelligence   Action quality assessment   long-term action quality assessment   domain shift   flow matching  

摘要： Long-term Action Quality Assessment (AQA) aims to evaluate the quantitative performance of actions in long videos. However, existing methods face challenges due to domain shifts between the pre-trained large-scale action recognition backbones and the specific AQA task, thereby hindering their performance. This arises since fine-tuning resource-intensive backbones on small AQA datasets is impractical. We address this by identifying two levels of domain shift: task-level, regarding differences in task objectives, and feature-level, regarding differences in important features. For feature-level shifts, which are more detrimental, we propose Progressive Hierarchical Instruction (PHI) with two strategies. First, Gap Minimization Flow (GMF) leverages flow matching to progressively learn a fast flow path that reduces the domain gap between initial and desired features across shallow to deep layers. Additionally, a temporally-enhanced attention module captures long-range dependencies essential for AQA. Second, List-wise Contrastive Regularization (LCR) facilitates coarse-to-fine alignment by comprehensively comparing batch pairs to learn fine-grained cues while mitigating domain shift. Integrating these modules, PHI offers an effective solution. Experiments demonstrate that PHI achieves state-of-the-art performance on three representative long-term AQA datasets, proving its superiority in addressing the domain shift for long-term AQA.

关键词： Genomics   Bioinformatics   Sequential analysis   Computer security   Security   Data privacy   DNA   Computer science   Biotechnology   Information systems   Next generation sequencing (NGS)   cyber-biosecurity   genomic data security   malware attack   genomic data privacy   cyber attacks on NGS  

摘要： Next-generation sequencing (NGS) has transformed genomic research and healthcare by enabling the rapid and cost-effective sequencing of DNA and RNA, surpassing traditional techniques such as Sanger sequencing. This technological leap has had a profound impact on fields including biomedical research, personalised medicine, cancer genomics, agriculture, and forensic sciences. With its widespread adoption, NGS has made genomic information more accessible, facilitating the sequencing of millions of genomes. However, the growing reliance on NGS has also brought significant challenges related to cyber-biosecurity, particularly the protection of genomic data against cyber threats such as unauthorised access, data breaches, and exploitation. Genomic data is inherently sensitive, and vulnerabilities in NGS technologies, software, data-sharing practices, and open-access databases expose it to risks concerning data confidentiality, integrity, and privacy. While NGS data plays an indispensable role across numerous sectors, research addressing the cyber-biosecurity of these technologies remains fragmented. Most existing studies focus narrowly on specific areas, such as microbial sequencing or system architecture, and fail to provide a holistic perspective on the security challenges that span the entire NGS workflow. Additionally, the lack of interdisciplinary collaboration between the biotechnology and cybersecurity communities further exacerbates these gaps. This paper seeks to bridge these gaps by thoroughly examining cyber-biosecurity threats throughout the NGS workflow. It introduces a tailored taxonomy specifically designed for NGS, aimed at increasing stakeholder awareness of potential vulnerabilities and threats. Key insights include identifying vulnerabilities at various stages of the NGS process-from data generation to analysis and storage-and categorising these threats systematically. The study highlights critical gaps in current research, underscoring the need for int

关键词： Computer Technology   Railway Transportation   Technological Innovation   Technology Transfer   MPSO Algorithm  

摘要： This article aims to discuss the influence of computer technology on the innovation and technology transfer of modern railway transportation. By historical analysis and empirical research, this article first reviews the evolution of railway transportation from steam locomotive to High-Speed Railways, and then compares and analyzes the performance of the improved multi-objective particle swarm optimization (MPSO) algorithm and traditional particle swarm optimization (PSO) algorithm in train operation scheduling through designing simulation experiments. The experiment selected the representative train operation data sets, and the first 25 and 15 data items are used as training sets respectively to predict the subsequent data items, and the prediction results of MPSO algorithm, traditional PSO algorithm and support vector machine algorithm are compared. The results show that MPSO algorithm is excellent in prediction accuracy, convergence speed and error control, and its prediction results are closer to the actual train running state, which verifies the effectiveness of computer technology in optimizing railway transportation scheduling.

关键词： 区块链技术   高校计算机专业   教学资源集成   教学资源共享   资源存储  

摘要： 高校计算机专业教学资源共享时，版权保护技术薄弱，资源易被侵权，缺乏追踪限制机制，影响教师积极性。为解决此问题，文章提出基于区块链技术的集成共享方法：构建集成框架，用面向服务架构和Web Services技术连接整合各教学资源系统；设计基于区块链的安全存储与防篡改机制，利用去中心化特性、Merkle树结构和SHA-256哈希函数加密确保资源安全存储；量化资源相似程度，结合传输效率计算与传输函数实现资源共享。实验表明，该方法借区块链去中心化存储特性提升共享速度，提高资源共享效率。

关键词： Droughts   Machine learning   Asia   Deep learning   Indexes   Water resources   Remote sensing   Predictive models   Computer science   Climate change   Environmental monitoring   Economics   Drought modeling   drought prediction   machine learning   deep learning   South Asia  

摘要： Drought is an environmental and economic problem. Sustainable ecosystems, water resources, food security, and ecosystem sustainability. Machine all are severely affected by drought. Due to the increasing frequency and severity of droughts caused by climate change. Effective drought modeling is crucial for early warning systems and risk mitigation. Recent advances in machine learning (ML) and deep learning (DL) techniques have been developed as potential drought modeling tools, which offer accurate and reliable drought detection. This review paper summarizes the drought modeling(Drought Prediction, Drought Detection and Drought Forecasting) approaches. This paper focuses on three main aspect. 1) The selection of the region for this study, for this study South Asia(SA) is selected as region of interest (ROI) that offer accurate drought modeling, providing policymakers and decision-makers with insightful information. The geographical scope of this study is the region of South Asia. This region is selected because of its heavy reliance on agriculture. 2) This paper focuses on the current and future trends, challenges, and advances of and vulnerability to droughts. The review offers a thorough grasp of how drought conditions are evaluated by gathering and analyzing the most important drought indicators and metrics specific to South Asia. The paper explores the current state-of-the-art in ML and DL for drought modeling. 3) This review encapsulates the indicator and metrics (Complex Machine learning and deep learning models) for drought modeling which are most relevant to the SA region. This study sum up as most common challenges in drought modeling are, highlighting current challenges such as incomplete and inconsistent datasets, lack of explainable and interpretable models, and unavailability of data for model uncertainty analysis. This study proposes that these problems can be solved with modern machine learning techniques such as explainable machine learning and federa

关键词： Videos   Visualization   Semantics   Chatbots   Image recognition   Electronic mail   Refining   Computer science   Benchmark testing   Training   Egocentric Action Recognition   Zero-Shot Learning   Vision-Language Learning  

摘要： Vision-Language Models (VLMs), pre-trained on large-scale datasets, have shown impressive performance in various visual recognition tasks. This advancement paves the way for notable performance in some egocentric tasks, Zero-Shot Egocentric Action Recognition (ZS-EAR), entailing VLMs zero-shot to recognize actions from first-person videos enriched in more realistic human-environment interactions. Typically, VLMs handle ZS-EAR as a global video-text matching task, which often leads to suboptimal alignment of vision and linguistic knowledge. We propose a refined approach for ZS-EAR using VLMs, emphasizing fine-grained concept-description alignment that capitalizes on the rich semantic and contextual details in egocentric videos. In this work, we introduce a straightforward yet remarkably potent VLM framework, aka GPT4Ego, designed to enhance the fine-grained alignment of concept and description between vision and language. Specifically, we first propose a new Ego-oriented Text Prompting (EgoTP ) scheme, which effectively prompts action-related text-contextual semantics by evolving word-level class names to sentence-level contextual descriptions by ChatGPT with well-designed chain-of-thought textual prompts. Moreover, we design a new Ego-oriented Visual Parsing (EgoVP ) strategy that learns action-related vision-contextual semantics by refining global-level images to part-level contextual concepts with the help of SAM. Extensive experiments demonstrate GPT4Ego significantly outperforms existing VLMs on three large-scale egocentric video benchmarks, i.e., EPIC-KITCHENS-100 (33.2% up arrow (+9.4) ), EGTEA (39.6% up arrow (+5.5) ), and CharadesEgo (31.5% up arrow (+2.6) ). In addition, benefiting from the novel mechanism of fine-grained concept and description alignment, GPT4Ego can sustainably evolve with the advancement of ever-growing pre-trained foundational models. We hope this work can encourage the egocentric community to build more investigation into pre-trained v

关键词： Social networking (online)   Computational modeling   Collective intelligence   Multi-agent systems   Receivers   Resistance   Mathematical models   Hands   Drones   Computer science   Agent based modelling   computational collective intelligence   social collective   consensus theory   group interaction  

摘要： This paper presents a comprehensive description of the social collective, an approach to group modeling for computational collective intelligence purposes. The social collective is independent of the structure of connections between members of the group, therefore it may use any existing representation of a social graph as a basis. Its modular form describes the internal mechanisms of group members: what type of data or knowledge do they use, how is it modified based on internal factors and on communication, as well as how communication between members works. A measure of drift is described in several variants, to better represent the behavior of the social collective. The measure is derived from behavior of social groups and past multi-agent models. Simulation experiments are used to show the observable effects in terms of drift, and to compare to non-collective approaches to group modeling. Social collective is specialized for computational collective intelligence tasks, but it may be fine-tuned for classical tasks, e.g., knowledge diffusion.

关键词： Urban areas   Routing   Benchmark testing   Metaheuristics   Genetic algorithms   Vehicle routing   Surface waves   Standards   Heuristic algorithms   Computer science   Travelling salesperson problem (TSP)   vehicle routing problem (VRP)   nature inspired algorithms   transverse waves  

摘要： This study introduces the Transverse Waves Algorithm (TWA), a novel heuristic approach devised to address the Travelling Salesperson Problem (TSP) within a unique instance. Inspired by natural phenomena where energy disperses in waves from a central point and interacts with its environment, TWA is designed for areas structured in concentric circles and segmented into equal parts. The algorithm mimics wave behaviors, when disrupted by obstacles and reflected by surfaces, to optimize routing. It systematically calculates routes initiating from a central hub, extending along each circle's circumference, visiting each location once, and progressing through strategically chosen adjacent points to minimize the total traveled distance. Experimental results indicate that TWA significantly enhances route optimization, outperforming traditional methods in specific instances. The findings underscore TWA's potential contributions to the field of logistics and operational research.

关键词： Electronic commerce   Consumer behavior   Organisms   Technological innovation   Systematic literature review   Psychology   Mathematical models   Computer science   Transforms   Stakeholders   e-commerce   cognitive   affective   SOR theory   Indonesia  

摘要： Technology advancement has given rise to a new online purchasing paradigm known as e-commerce. Online shopping in e-commerce platform has altered consumer behavior and purchasing decisions. However, a few studies have conducted by considering cognitive-affective process in making purchase decision on e-commerce. Therefore, the current study examines consumer behavior from perspective of stimulus-organism-response theory and the internal perception process (cognitive and affective) to understand consumer's purchase decisions. We use Structural Equation Model-Partial Least Square technique to analyze the data of 308 consumers in Indonesia. Our findings report that two dimensions of the stimulus environment (credibility and product usefulness) significantly contributed to perceived trust;convenience has an insignificant impact on perceived trust and perceived value. In comparison, three dimensions of the stimulus environment (convenience, informativeness, and product price) contribute to consumers' perceived trust. Furthermore, the stimulus environment indicators (credibility, informativeness, product price, and product usefulness) effect on perceived value, while convenience has an insignificant on consumers' perceived value. The internal perception process or cognitive and affective process, perceived trust has an insignificant impact on perceived pleasure, while perceived value has a significant impact on perceived pleasure. Lastly, perceived pleasure significantly influences consumer purchase decisions in online shopping and e-commerce. The findings of this study can expand the research on e-commerce how consumer's consider their cognitive and affective factor on product purchase decision. The study suggests that the optimize the features and products as way to improve the consumer purchase decision, and it affects toward cognitive and affective components such as perceive trust, value, and pleasure. This study has contributed both theoretically and practically to

关键词： Protocols   Cryptography   Fault tolerant systems   Fault tolerance   Encryption   Public key   Forensics   Consensus protocol   Computer science   Training   Distributed consensus   Byzantine fault tolerance   timed-release operations   threshold public key encryption  

摘要： This paper introduces the Deferred Byzantine Generals Problem, a variant of the Byzantine Generals Problem which focuses on ensuring replicas maintain consistency over timed-release secret operations (operations that can only be known after a specified time or event). The solution to the problem is called the Deferred Byzantine Fault Tolerant (DBFT) consensus. DBFT can operate exclusive or be interleave with BFTs to handle specific tasks at designated sequence numbers or views, thereby facilitating the implementation of certain system-desirable features or supporting novel applications. It does not rely on existing timed-release primitives, but instead ensures its timed-release property through voting interactions. We presents the system model of DBFT SMR under partial synchronization using Threshold Public Key Encryption (TPKE) as the cryptographic primitives, highlighting the core issues. Then we design and implement the DBFT protocol using PBFT notations, focusing on the unique parts to facilitate expansions to other paradigms. Through experimental results, we show the impact of different executing modes and parameter choices on performance and discuss potential optimizations.