检索结果-内蒙古大学图书馆

TSCompiler: efficient compilation framework for dynamic-shape models

science China(Information sciences) 2024年第10期67卷 67-84页

作者： Xiang LUO Chen ZHANG Chenbo GENG Yanzhi YI Jiahui HU Renwei ZHANG Zhen ZHANG Gianpietro CONSOLARO Fan YANG Tun LU Ning GU Li SHANG School of Computer Science Fudan University School of Electronic Information and Electrical Engineering Shanghai Jiao Tong University School of Computer Science and Technology Harbin Institute of Technology Huawei Technologies Co. Ltd. Huawei Paris Research Center School of Microelectronics Fudan University

Today's deep learning models face an increasing demand to handle dynamic shape tensors and computation whose shape information remains unknown at compile time and varies in a nearly infinite range at runtime. This shape dynamism brings tremendous challenges for existing compilation pipelines designed for static models which optimize tensor programs relying on exact shape values. This paper presents TSCompiler, an end-to-end compilation framework for dynamic shape models. TSCompiler first proposes a symbolic shape propagation algorithm to recover symbolic shape information at compile time to enable subsequent optimizations. TSCompiler then partitions the shape-annotated computation graph into multiple subgraphs and fine-tunes the backbone operators from the subgraph within a hardware-aligned search space to find a collection of high-performance schedules. TSCompiler can propagate the explored backbone schedule to other fusion groups within the same subgraph to generate a set of parameterized tensor programs for fused cases based on dependence analysis. At runtime, TSCompiler utilizes an occupancy-targeted cost model to select from pre-compiled tensor programs for varied tensor shapes. Extensive evaluations show that TSCompiler can achieve state-of-the-art speedups for dynamic shape models. For example, we can improve kernel efficiency by up to 3.97× on NVIDIA RTX3090, and 10.30× on NVIDIA A100 and achieve up to five orders of magnitude speedups on end-to-end latency.

关键词： machine learning tensor compilers dynamic shape operator fusion code generation auto-tuning

来源：评论

学校读者我要写书评

暂无评论

Research on stock trend prediction method based on optimized random forest

引用

CAAI Transactions on Intelligence technology 2023年第1期8卷 274-284页

作者： Lili Yin Benling Li Peng Li Rubo Zhang School of Computer and Science and Technology Harbin University of Science and TechnologyHarbinChina Department of Computer Science and Engineering Dalian Nationalities UniversityDalianChina

As a complex hot problem in the financial field,stock trend forecasting uses a large amount of data and many related indicators;hence it is difficult to obtain sustainable and effective results only by relying on empirical *** in the field of machine learning have proved that random forest can form better judgements on this kind of problem,and it has an auxiliary role in the prediction of stock *** study uses historical trading data of four listed companies in the USA stock market,and the purpose of this study is to improve the performance of random forest model in medium-and long-term stock trend *** study applies the exponential smoothing method to process the initial data,calculates the relevant technical indicators as the characteristics to be selected,and proposes the D-RF-RS method to optimize random *** the random forest is an ensemble learning model and is closely related to decision tree,D-RF-RS method uses a decision tree to screen the importance of features,and obtains the effective strong feature set of the model as ***,the parameter combination of the model is optimized through random parameter *** experimental results show that the average accuracy of random forest is increased by 0.17 after the above process optimization,which is 0.18 higher than the average accuracy of light gradient boosting machine *** with the performance of the ROC curve and Precision–Recall curve,the stability of the model is also guaranteed,which further demonstrates the advantages of random forest in medium-and long-term trend prediction of the stock market.

关键词： ensemble learning finance random forest random search technical indicator

来源：评论

学校读者我要写书评

暂无评论

A novel method using matrix coding on-chain and sharing multimedia data for improved usability and reliability

引用

Multimedia Tools and Applications 2024年第40期83卷 87727-87748页

作者： Yang, Lizhu Qin, Yong School of Computer Science and Engineering Guangzhou Institute of Science and Technology Guangzhou China School of Cyberspace Security Dongguan University of Technology Dongguan China

Blockchain technology has the characteristics of non-tampering and forgery, traceability, and so on, which have good application advantages for the storage of multimedia data. So we propose a novel method using matrix coding on-chain and sharing multimedia data for improved usability and reliability. Based on matrix code, we provide the block matrix coding-based on-chain storing method and the block invertible matrix decoding-based sharing method. The method progressively converts the block to create linearly coupled coded chunks, so the blockchain only sends the chunk sets that have been encoded. Each node also maintains just the ledger relevant to its own operations, which lessens the node’s storage burden. Meanwhile, we make several replications of the chunk set by adding a replication factor to increase the feasibility of chunks. Only when all of the target nodes fail is decoding reconstruction necessary, which further enhances the read performance of the blockchain. Many experimental tests are conducted to evaluate the performance based on various parameters such as time overhead, storage overhead, compression factor, failure factor, and so on. According to theoretical analysis and experimental verification, the method offers good read performance with a high recovery success rate and minimal storage occupation, while guaranteeing the availability and dependability of the block data. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Blockchain

来源：评论

学校读者我要写书评

暂无评论

An efficient multi-objective task scheduling in edge computing using adaptive honey badger optimisation

引用

International Journal of Web engineering and technology 2024年第2期19卷 110-126页

作者： Nagalakshmi, Bantupalli Subramanian, Sumathy School of Computer Science and Engineering Vellore Institute of Technology Tamil Nadu Vellore632014 India School of Computer Science Engineering and Information Systems Vellore Institute of Technology Tamil Nadu Vellore632014 India

Task scheduling, which is important in cloud computing, is one of the most challenging issues in this area. Hence, an efficient and reliable task scheduling approach is needed to produce more efficient resource employment. So, a multi-objective-based task scheduling for edge computing is suggested in this study. This paper develops the adaptive honey badger optimisation algorithm (AHBA) to accomplish this goal. The lack of population, the original honey badger algorithm (HBO) has the issue of becoming trapped in local optima. To maintain population variety and improve convergence towards the ideal solution, HBO is combined with the opposition-based learning technique (OBL). Based on makespan, cost, energy consumption, and resource usage, the multi-objective function is created. According to simulation results, the proposed approach has a lot of potential in this field. Java and cloud Simulator are used to implement the suggested model. Copyright © 2024 Inderscience Enterprises Ltd.

关键词： Energy utilization

来源：评论

学校读者我要写书评

暂无评论

Automatic summarization of cooking videos using transfer learning and transformer-based models

引用

Discover Artificial Intelligence 2025年第1期5卷 1-20页

作者： Sadique, P. M. Alen Aswiga, R.V. School of Computer Science and Engineering Vellore Institute of Technology Tamil Nadu Chennai600127 India

The proliferation of cooking videos on the internet these days necessitates the conversion of these lengthy video contents into concise text recipes. Many online platforms now have a large number of cooking videos, in which, there is a challenge for viewers to extract comprehensive recipes from lengthy visual content. Effective summary is necessary in order to translate the abundance of culinary knowledge found in videos into text recipes that are easy to read and follow. This will make the cooking process easier for individuals who are searching for precise step by step cooking instructions. Such a system satisfies the needs of a broad spectrum of learners while also improving accessibility and user simplicity. As there is a growing need for easy-to-follow recipes made from cooking videos, researchers are looking on the process of automated summarization using advanced techniques. One such approach is presented in our work, which combines simple image-based models, audio processing, and GPT-based models to create a system that makes it easier to turn long culinary videos into in-depth recipe texts. A systematic workflow is adopted in order to achieve the objective. Initially, Focus is given for frame summary generation which employs a combination of two convolutional neural networks and a GPT-based model. A pre-trained CNN model called Inception-V3 is fine-tuned with food image dataset for dish recognition and another custom-made CNN is built with ingredient images for ingredient recognition. Then a GPT based model is used to combine the results produced by the two CNN models which will give us the frame summary in the desired format. Subsequently, Audio summary generation is tackled by performing Speech-to-text functionality in python. A GPT-based model is then used to generate a summary of the resulting textual representation of audio in our desired format. Finally, to refine the summaries obtained from visual and auditory content, Another GPT-based model is used

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

IensNet: A novel and efficient approach for iris spoof detection via ensemble of deep models

引用

Multimedia Tools and Applications 2025年 1-30页

作者： Sharma, Deepika Selwal, Arvind School of Computer Science Engineering and Technology Bennett University Greater Noida201310 India Department of Computer Science and Information Technology Central University of Jammu Samba India

Iris biometrics allow contactless authentication, which makes it widely deployed human recognition mechanisms since the couple of years. Susceptibility of iris identification systems remains a challenging task due to diversity in spoof or presentation attacks (PAs) that fails to assure consistency while adopting them in real life scenarios. Hence, iris PAs are the growing concerns that gained significant attention in recent past decade. To alleviate these attacks or recognize presentation attack instruments (PAIs), iris presentation attacks detection (IPAD) algorithms are designed to distinguish a real and fabricated iris trait. Aiming at the efficient iris spoof detection mechanism, in this research work we expound a novel ensemble learning-enabled model (IensNet) that learns three pre-trained and fined-tuned deep models (i.e. DenseNet161, ResNet and VGGNet) for better accuracy and generalized performance. The novel IensNet approach offers several merits (i.e. consolidated strengths of multiple models, improved generalization ability, etc.) as compared to a simple transfer learning strategy where the knowledge is drawn from single pre-trained model. Finally, our approach learns a novel fully-connected dual layer classifier via outcome of three fine-tuned models to yield a final classification result as bonafide or spoof iris trait. Our approach is evaluated on Notre Dame LivDet iris 2017 and Notre Dame contact lenses 2015 anti-spoofing datasets. The experimental analysis of IensNet offers outstanding performance with a lower ACER of 0.2% and 1.4% for Iris-LivDet-2017 and Notre Dame contact lenses 2015 dataset respectively. Besides, IensNet exhibit promising results in cross-dataset environment with an ACA of 91.46%. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Advancing differential diagnosis: a comprehensive review of deep learning approaches for differentiating tuberculosis, pneumonia, and COVID-19

引用

Multimedia Tools and Applications 2025年第13期84卷 11871-11906页

作者： Kansal, Kajal Chandra, Tej Bahadur Singh, Akansha School of Computer Science Engineering and Technology Bennett University Uttar Pradesh Greater Noida India

In the realm of medical diagnostics, particularly in differential diagnosis, where differentiating between illnesses or ailments with comparable symptoms is essential, deep learning has gained importance. Recent developments in deep learning have demonstrated considerable promise for revolutionizing medical diagnostics by using the ability of artificial intelligence (AI) to accurately interpret radiological images. We examine the most cutting-edge deep learning techniques currently being utilized for the differential diagnosis of tuberculosis, pneumonia, and COVID-19 in this in-depth review. The study presents an in-depth critical review of several SOTA (state-of-the-art) studies used for differential diagnosis of different respiratory abnormalities like TB, Pneumonia, and COVID-19. In addition, an overview of various approaches, datasets employed in each method, various diagnosis tests, used assessment measures, and obtained performance is summarized and comprehensively compared to assist future research. We suggest a pathway for future research and development of deep learning solutions for differential diagnosis by critically analyzing the current literature and outlining the limitations and potential in this sector. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： COVID-19

来源：评论

学校读者我要写书评

暂无评论

Leveraging Concise Concepts With Probabilistic Modeling for Interpretable Visual Recognition

引用

IEEE Transactions on Multimedia 2025年 27卷 3117-3131页

作者： Zhang, Yixuan Liu, Chuanbin Liu, Yizhi Gao, Yifan Lu, Zhiying Xie, Hongtao Zhang, Yongdong University of Science and Technology of China School of Information Science and Technology Hefei230027 China Hunan University of Science and Technology Department of Computer Science and Engineering Hunan411199 China

Interpretable visual recognition is essential for decision-making in high-stakes situations. Recent advancements have automated the construction of interpretable models by leveraging Visual Language Models (VLMs) and Large Language Models (LLMs) with Concept Bottleneck Models (CBMs), which process a bottleneck layer associated with human-understandable concepts. However, existing methods suffer from two main problems: a) the collected concepts from LLMs could be redundant with task-irrelevant descriptions, resulting in an inferior concept space with potential mismatch. b) VLMs directly map the global deterministic image embeddings with fine-grained concepts results in an ambiguous process with imprecise mapping results. To address the above two issues, we propose a novel solution for CBMs with Concise Concept and Probabilistic Modeling (CCPM) that can achieve superior classification performance via high-quality concepts and precise mapping strategy. First, we leverage in-context examples as category-related clues to guide LLM concept generation process. To mitigate redundancy in the concept space, we propose a Relation-Aware Selection (RAS) module to obtain a concise concept set that is discriminative and relevant based on image-concept and inter-concept relationships. Second, for precise mapping, we employ a Probabilistic Distribution Adapter (PDA) that estimates the inherent ambiguity of the image embeddings of pre-trained VLMs to capture the complex relationships with concepts. Extensive experiments indicate that our model achieves state-of-the-art results with a 6.18% improvement in classification accuracy on eight mainstream recognition benchmarks as well as reliable explainability through interpretable analysis. © 1999-2012 IEEE.

关键词： Decision making

来源：评论

学校读者我要写书评

暂无评论

Modifying Lyapunov exponent of chaotic map by self-cascading

引用

science China(Technological sciences) 2024年第7期67卷 2203-2214页

作者： YI ChenLong LI ChunBiao LI YongXin XIA Ming HUA ZhongYun School of Electronic and Information Engineering Nanjing University of Information Science&TechnologyNanjing 210044China School of Artificial Intelligence Nanjing University of Information Science&TechnologyNanjing 210044China School of Computer Science and Technology Harbin Institute of Technology Shenzhen Graduate SchoolShenzhen 518055China

The self-cascade(SC) method is an effective technique for chaos enhancement and complexity increasing in chaos ***, the controllable self-cascade(CSC) method allows for more accurate control of Lyapunov exponents of the discrete map. In this work, the SC and CSC systems of the original map are derived, which enhance the chaotic performance while preserving the fundamental dynamical characteristics of the original map. Higher Lyapunov exponent of chaotic sequences corresponding to higher frequency are obtained in SC and CSC systems. Meanwhile, the Lyapunov exponent could be linearly controlled with greater flexibility in the CSC system. The verification of the numerical simulation and theoretical analysis is carried out based on the platform of CH32.

关键词： self-cascade controllable self-cascade chaos enhancement chaos control

来源：评论

学校读者我要写书评

暂无评论

A Neural Network Slab Quality Prediction Analysis Based on a Hybrid Intelligent Optimisation Algorithm

IAENG International Journal of Computer Science

引用

IAENG International Journal of computer science 2025年第3期52卷 673-683页

作者： Li, Yiran Zhang, Chunna School of Applied Technology University of Science and Technology Liaoning Liaoning Anshan114051 China School of Computer Science and Software Engineering University of Science and Technology Liaoning Liaoning Anshan114051 China

To address the problem of inaccurate prediction of slab quality in continuous casting, an algorithm based on particle swarm optimisation and differential evolution is proposed. The algorithm combines BP neural network prediction method. Firstly, the factors affecting the quality of the slab are analysed and a sufficient number of sample sets are extracted for training;Secondly, the particle swarm optimisation algorithm is optimised, and then the BP neural network is optimised and the chaos mechanism is introduced to increase the continuity of motion of the particles. The population is partitioned hierarchically, and different types of particles are processed separately to improve the global search capability of the algorithm. The differential evolution algorithm is also integrated to optimise cross-selection and increase particle diversity. Finally, a prediction model is built to predict and analyse the sample data. In the experiment, three algorithms are tested for different benchmark functions, namely the comparison of convergence, stability and searchability. The optimisation time comparison is completed, the target curve fit test is performed and the slab quality test results are checked. The experimental results show that the PSO-EIDE algorithm has strong optimisation ability, fast convergence speed and obvious improvement in stability and accuracy. © (2025), (International Association of Engineers). All rights reserved.

关键词： Prediction models

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：