检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Sun, Wenhao Tu, Rong-Cheng Liao, Jingyi Jin, Zhao Tao, Dacheng College of Computing and Data Science Nanyang Technological University Singapore Singapore

Diffusion Transformers (DiTs) have proven effective in generating high-quality videos but are hindered by high computational costs. Existing video DiT sampling acceleration methods often rely on costly fine-tuning or exhibit limited generalization capabilities. We propose Asymmetric Reduction and Restoration (AsymRnR), a training-free and model-agnostic method to accelerate video DiTs. It builds on the observation that redundancies of feature tokens in DiTs vary significantly across different model blocks, denoising steps, and feature types. Our AsymRnR asymmetrically reduces redundant tokens in the attention operation, achieving acceleration with negligible degradation in output quality and, in some cases, even improving it. We also tailored a reduction sched- © 2024, CC BY.

关键词： Acceleration

来源：评论

学校读者我要写书评

暂无评论

Deep Convolutional Neural Network for Enhancing Traffic Sign Recognition Developed on Yolo V5

Deep Convolutional Neural Network for Enhancing Traffic Sign...

引用

Advancements in Smart, Secure and Intelligent computing (ASSIC), International Conference on

作者： Aofan Liu Yutong Liu Saif Kifah School of Computing and Data Science Xiamen University Malaysia Sepang Malaysia

In today's era, deep learning neural networks with multiple hidden layers have been widely used in many fields. The deep learning method has more powerful features that enhance the method's performance by a learning process. With the development of the logistics industry and the prevalence of autonomous driving, traffic sign recognition has gained rising attention. This study uses a YOLO CNN to classify traffic signs. To improve model performance, we used MSRCR image augmentation during preprocessing. In the improvement phase, we used YOLOv5 to automate traffic sign categorization and improved training methods and network architecture. GTSRB and CCTSDB were used to assess the proposed technique. The experimental results show that the YOLOv5 model outperforms other methods. It has a 99.8% accuracy rate in the GTSRB dataset and 98.4% precision in the CCTSDB.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Dual-Port, and Single-Element Planar Inverted-F Antenna for Current 4G/5G Portable Applications 21

A Dual-Port, and Single-Element Planar Inverted-F Antenna fo...

引用

21st International Multi-Conference on Systems, Signals and Devices, SSD 2024

作者： Alja'Afreh, Saqer S. Altakhaineh, Amjaad T. Al-Shamaileh, Mallak H. Matarneh, Aser M Daoud, Omar R. Al-Khawaldah I, Mohammad School of Engineering and Computing American International University Electrical and Computing Engineering Al-Jahra Kuwait Pioneers International Academy Data Science and Machine Learning Dept. Amman Jordan Al-Ahliyya Amman University Communications and Computer Engineering Amman Jordan

ISBN: (纸本)9798350374131

A first-hand, low-profile, two-port planar inverted-F antenna is proposed for current 4G/5G sub-3 GHz hand-portable multiple-input and multiple-output (MIMO) applications. The design features a profile of about 5 mm and dual-coplanar feeds, which are isolated using only a modified top plate, while the shared ground plane is not perturbed as in previous designs. An open-edge slot on the top of the PIFA provides a 3λ/4 current path. The proposed antenna covers approximately a 15% impedance bandwidth fraction in the range of 2.45-2.85 GHz. The simulated results, including scattering parameters, far-field radiation patterns, and the envelope correlation coefficient, all exhibit excellent features. © 2024 IEEE.

关键词： Antenna feeders

来源：评论

学校读者我要写书评

暂无评论

Detecting Criminal Activities and Promoting Safety Using Deep Learning

Detecting Criminal Activities and Promoting Safety Using Dee...

引用

2022 IEEE International Conference on Advances in computing, Communication and Applied Informatics, ACCAI 2022

作者： Mathur, Rohan Chintala, Tejas Rajeswari, D. School of Computing SRM Institute of Science and Technology Department of Data Science and Business Systems Tamil Nadu 603203 India

ISBN: (纸本)9781665495295

Automation and autonomous systems are among the few powerhouses of innovation that drive entire domains towards advancing further in leaps and bounds. Great technological innovations can be attributed to tasks that are made easier and more perceptible by automation, and artificial intelligence is here to make these automated systems smart enough to perform their tasks with the power of decision-making, thereby greatly reducing human intervention in redundant processes. Our project follows the aforementioned ideals: building a product to minimize manual labor (both physical and mental) for tasks that can be seamlessly automated and processed while solving the main problem statement at hand. Currently, surveillance cameras play a vital role to ensure the safety of the people, yet they are plain video-providing entities with no smart decision making mechanisms of their own. Because of this growth of data composed from surveillance cameras, automated video streams have become a requisite for automatically detecting abnormal events. The main aim of the project focuses on promoting safety on campus by employing deep learning techniques to automate the task of monitoring and reporting crimes from the physical Closed-Circuit Television (CCTV), assigning the responsibility of detecting criminal activity to a framework that can identify patterns to differentiate them for smarter monitoring. The model in this paper that we propose was able to distinguish between certain crimes with a 0.94 and 0.95 precision for Assault and Abuse respectively. © 2022 IEEE.

关键词： Crime

来源：评论

学校读者我要写书评

暂无评论

Deep Neural Network Approach on Phishing Websites Detection, Identification and Proactive Prevention 8

Deep Neural Network Approach on Phishing Websites Detection,...

引用

8th IEEE International Conference on Computational System and Information Technology for Sustainable Solutions, CSITSS 2024

作者： Soundarya, K. Siddaraju, H.S. Tekchandani, Hitesh Bhardwaj, Navodit Rani, Pooja Joshi, Pallavi Amrita School of Computing Amrita Vishwa Vidyapeetham Department of Computer Science Mysuru India Koneru Lakshmaiah Education Foundation Artificial Intelligence and Data Science Hyderabad India Researcher at Food and Drug Administration MD United States

ISBN: (纸本)9798331505462

Phishing is a prevalent type of cyberattack that involves posing as a trustworthy source in an email message to fraudulently attempt to get personal information such as usernames, passwords, and bank account information. Strong detection and prevention methods are required to address the phishing issue. The objective is to develop efficient algorithms that use machine learning (ML) and data mining methods to accurately detect and categorize phishing URLs. This initiative aims to support ongoing initiatives to strengthen cybersecurity and shield consumers from the risks associated with phishing scams. It can be observed from the results that the Deep neural network model gives 95% of accuracy and 96% of precision for classification and detection of phishing websites. © 2024 IEEE.

关键词： Phishing

来源：评论

学校读者我要写书评

暂无评论

Towards Responsible and Ethical AI Chatbot in Education: A Guideline for Developers

Towards Responsible and Ethical AI Chatbot in Education: A G...

引用

Robotics, Automation and Artificial Intelligence (RAAI), IEEE International Conference on

作者： Jiayun Mo Ong Chin Ann College of Computing and Data Science Nanyang Technological University Singapore Singapore

ISBN: (数字)9798331520038

ISBN: (纸本)9798331520045

With the expansion of generative Artificial Intelligence (AI) and the penetration of AI chatbots into every aspect of people's daily lives, discussions and issues related to AI ethics and responsibility began to arise simultaneously. A large portion of AI users remain ignorant of such ethical discussions, while a significant percentage of developers lack precautionary measures to prevent irresponsible AI behaviors. An increasing number of companies and organizations are taking action to address the problems. Despite the efforts, prominent issues still prevail. Most existing solutions are too general and theoretical; few are narrowed down to examine the specific application scenario of AI chatbots in education. In response, this study aims to address the arising concerns and the lack of a satisfying solution. After conducting an extensive literature review and performing a detailed analysis of survey data, this study develops a comprehensive guideline about AI ethics and responsible AI, which specifically targets undergraduate developers and guides them during their development of AI chatbots used in education. This guideline successfully bridges the knowledge gap, raises awareness about AI ethics, and acts as the first guideline for undergraduate students, influencing future development practices.

关键词： Surveys Ethics Systematics Generative AI Educational robots Education Chatbots Artificial intelligence Guidelines Systematic literature review

来源：评论

学校读者我要写书评

暂无评论

Plausible Heterogeneous Graph k-Anonymization for Social Networks

引用

Tsinghua science and Technology 2022年第6期27卷 912-924页

作者： Kaiyang Li Ling Tian Xu Zheng Bei Hui School of Computer Science and Engineering University of Electronic Science and Technology of ChinaChengdu 611731China School of Information and Software Engineering University of Electronic Science and Technology of ChinaChengdu 610054China Trusted Cloud Computing and Big Data Key Laboratory of Sichuan Province Chengdu 610000China

The inefficient utilization of ubiquitous graph data with combinatorial structures necessitates graph embedding methods,aiming at learning a continuous vector space for the graph which is amenable to be adopted in traditional machine learning algorithms in favor of vector *** embedding methods build an important bridge between social network analysis and data analytics as social networks naturally generate an unprecedented volume of graph data *** social network data not only bring benefit for public health,disaster response,commercial promotion,and many other applications,but also give birth to threats that jeopardize each individual’s privacy and ***,most existing works in publishing social graph embedding data only focus on preserving social graph structure with less attention paid to the privacy issues inherited from social *** be specific,attackers can infer the presence of a sensitive relationship between two individuals by training a predictive model with the exposed social network *** this paper,we propose a novel link-privacy preserved graph embedding framework using adversarial learning,which can reduce adversary’s prediction accuracy on sensitive links while persevering sufficient non-sensitive information such as graph topology and node attributes in graph *** experiments are conducted to evaluate the proposed framework using ground truth social network datasets.

关键词： social network graph embedding privacy preservation adversarial learning

来源：评论

学校读者我要写书评

暂无评论

data Processing Model of Students' Evaluation of Teaching Based on Outlier Detection 5

Data Processing Model of Students' Evaluation of Teaching Ba...

引用

5th IEEE Eurasian Conference on Educational Innovation, ECEI 2022

作者： Qu, Ying Guo, Fei Hebei University of Science and Technology School of Economics and Management Data Science and Intelligent Computing Research Center Hebei China

ISBN: (纸本)9781665433181

The problem of low credibility of students in teaching evaluation is caused by ignoring non-teaching factors such as students' emotions, class culture, and course nature in the existing data of university students' teaching evaluation. A data processing model of students' teaching evaluation is constructed based on outlier detection. The outliers of the original data are screened and optimized based on students' emotional differences. Based on the analysis of non-teaching factors such as class culture and curriculum characteristics, the abnormal data of students' evaluation of teaching are processed. This improves the credibility of students' evaluation of teaching data to a certain extent. For the same sample data, the optimized data processing model of students' evaluation of teaching shows realistic evaluation results. The guidance of students' evaluation of teaching needs to be strengthened based on the original intention of teaching quality evaluation. It is suggested that evaluation models are scientifically selected, and the evaluation results are rationally used. © 2022 IEEE.

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

S2AC: ENERGY-BASED REINFORCEMENT LEARNING WITH STEIN SOFT ACTOR CRITIC 12

S2AC: ENERGY-BASED REINFORCEMENT LEARNING WITH STEIN SOFT AC...

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Messaoud, Safa Mokeddem, Billel Xue, Zhenghai Pang, Linsey An, Bo Chen, Haipeng Chawla, Sanjay Qatar Computing Research Institute Hamad Bin Khalifa University Qatar School of Computer Science and Engineering Nanyang Technological University Singapore SalesForce United States Skywork AI Data Science William & Mary United States

Learning expressive stochastic policies instead of deterministic ones has been proposed to achieve better stability, sample complexity, and robustness. Notably, in Maximum Entropy Reinforcement Learning (MaxEnt RL), the policy is modeled as an expressive Energy-Based Model (EBM) over the Q-values. However, this formulation requires the estimation of the entropy of such EBMs, which is an open problem. To address this, previous MaxEnt RL methods either implicitly estimate the entropy, resulting in high computational complexity and variance (SQL), or follow a variational inference procedure that fits simplified actor distributions (e.g., Gaussian) for tractability (SAC). We propose Stein Soft Actor-Critic (S2AC), a MaxEnt RL algorithm that learns expressive policies without compromising efficiency. Specifically, S2AC uses parameterized Stein Variational Gradient Descent (SVGD) as the underlying policy. We derive a closed-form expression of the entropy of such policies. Our formula is computationally efficient and only depends on first-order derivatives and vector products. Empirical results show that S2AC yields more optimal solutions to the MaxEnt objective than SQL and SAC in the multi-goal environment, and outperforms SAC and SQL on the MuJoCo benchmark. Our code is available at: https://***/SafaMessaoud/S2AC-Energy-Based-RL-with-Stein-Soft-Actor-Critic. © 2024 12th International Conference on Learning Representations, ICLR 2024. All rights reserved.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Improving VTE Identification through Language Models from Radiology Reports: A Comparative Study of Mamba, Phi-3 Mini, and BERT

Improving VTE Identification through Language Models from Ra...

引用

2024 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2024

作者： Deng, Jamie Wu, Yusen Yesha, Yelena Nguyen, Phuong University of Miami Department of Computer Science MiamiFL United States University of Miami Frost Institute for Data Science and Computing MiamiFL United States University of Miami Department of Radiology MiamiFL United States

ISBN: (纸本)9798350386226

Venous thromboembolism (VTE) is a critical cardio-vascular condition, encompassing deep vein thrombosis (DVT) and pulmonary embolism (PE). Accurate and timely identification of VTE is essential for effective medical care. This study builds upon our previous work, which addressed VTE detection using deep learning methods for DVT and a hybrid approach combining deep learning and rule-based classification for PE. Our earlier approaches, while effective, had two major limitations: they were complex and required expert involvement for feature engineering of the rule set. To overcome these challenges, we utilize the Mamba architecture-based classifier. This model achieves remarkable results, with a 97% accuracy and F1 score on the DVT dataset and a 98% accuracy and F1 score on the PE dataset. In contrast to the previous hybrid method on PE identification, the Mamba classifier eliminates the need for hand-engineered rules, significantly reducing model complexity while maintaining comparable performance. Additionally, we evaluated a lightweight Large Language Model (LLM), Phi-3 Mini, in detecting VTE. While this model delivers competitive results, outperforming the baseline BERT models, it proves to be computationally intensive due to its larger parameter set. Our evaluation shows that the Mamba-based model demonstrates superior performance and efficiency in VTE identification, offering an effective solution to the limitations of previous approaches. © 2024 IEEE.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：