检索结果-内蒙古大学图书馆

VeML: An End-to-End Machine Learning Lifecycle for Large-scale and High-dimensional data

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Le, Van-Duc Bui, Cuong-Tien Li, Wen-Syan Department of Electrical and Computer Engineering Seoul National University Korea Republic of Graduate School of Data Science Seoul National University Korea Republic of

An end-to-end machine learning (ML) lifecycle consists of many iterative processes, from data preparation and ML model design to model training and then deploying the trained model for inference. When building an end-to-end lifecycle for an ML problem, many ML pipelines must be designed and executed that produce a huge number of lifecycle versions. Therefore, this paper introduces VeML, a Version management system dedicated to end-to-end ML Lifecycle. Our system tackles several crucial problems that other systems have not solved. First, we address the high cost of building an ML lifecycle, especially for large-scale and high-dimensional dataset. We solve this problem by proposing to transfer the lifecycle of similar datasets managed in our system to the new training data. We design an algorithm based on the core set to compute similarity for large-scale, high-dimensional data efficiently. Another critical issue is the model accuracy degradation by the difference between training data and testing data during the ML lifetime, which leads to lifecycle rebuild. Our system helps to detect this mismatch without getting labeled data from testing data and rebuild the ML lifecycle for a new data version. To demonstrate our contributions, we conduct experiments on real-world, large-scale datasets of driving images and spatiotemporal sensor data and show promising results. © 2023, CC BY-NC-ND.

关键词： Machine learning

Multi-view Semantic Consistency based Information Bottleneck for Clustering

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Yan, Wenbiao Zhu, Jihua Zhou, Yiyang Wangr, Yifei Zheng, Qinghai School of Software Engineering Xi'an Jiaotong University Xi'An710049 China College of Computer and Data Science Fuzhou University Fuzhou350108 China

Multi-view clustering can make use of multi-source information for unsupervised clustering. Most existing methods focus on learning a fused representation matrix, while ignoring the influence of private information and noise. To address this limitation, we introduce a novel Multi-view Semantic Consistency based Information Bottleneck for clustering (MSCIB). Specifically, MSCIB pursues semantic consistency to improve the learning process of information bottleneck for different views. It conducts the alignment operation of multiple views in the semantic space and jointly achieves the valuable consistent information of multi-view data. In this way, the learned semantic consistency from multi-view data can improve the information bottleneck to more exactly distinguish the consistent information and learn a unified feature representation with more discriminative consistent information for clustering. Experiments on various types of multi-view datasets show that MSCIB achieves state-of-the-art performance. Copyright © 2023, The Authors. All rights reserved.

关键词： Semantics

Research on Algorithm of Building Overall Contour for Planar CAD Drawing

学校读者我要写书评

暂无评论

Research on Algorithm of Building Overall Contour for Planar...

Chinese Automation Congress (CAC)

作者： Hui Zhong Lei Fan Liaoning Province Big Data Management and Analysisline Laboratory of Urban Construction School of Computer Science and Engineering Shenyang Jianzhu University Shenyang China

In the study of automatic modeling of building climbing frame based on building CAD plans, recognizing the building's contour is a very critical step. In order to realize the automatic recognition of building contour and realize from 2D building drawings to 3D building entities, this paper generates 3D Autodesk Revit model from building plan AutoCAD 2D graphic recognition of building contour to realize BIM reconstruction. On the basis of adaptive chunking algorithm, combined with convex packet algorithm and depth search algorithm, an algorithm based on the recognition and processing of straight line Line and multi-segmented line PolyLine in architectural CAD drawings is proposed, and the building contour is automatically acquired and generated into a 3D entity. The algorithm links the CAD drawings in Revit, reads the text information by connecting to the. dwg database with the help of Teigha library, filters the straight lines and polylines of the target layer, processes them, and obtains and stores the endpoint information. The points on the building contour are obtained by scanning and comparing the endpoint coordinates. Finally, all the points on the contour are connected in an orderly manner to get the 2D contour of the building. Finally, the contour is solidly stretched in Revit to get the 3D building solid, which lays a certain foundation for the subsequent research on the automatic generation of the 3D model of the climbing frame. The algorithm is of great significance for realizing the automatic modeling from AutoCAD 2D graphics to Revit 3D model and the automatic modeling of building climbing frame, which has strong application value in the field of computer-aided building construction design.

关键词：

Theoretical Insights for Diffusion Guidance: A Case Study for Gaussian Mixture Models

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Wu, Yuchen Chen, Minshuo Li, Zihao Wang, Mengdi Wei, Yuting Department of Statistics and Data Science The Wharton School University of Pennsylvania United States Department of Electrical and Computer Engineering Princeton University United States

Diffusion models benefit from instillation of task-specific information into the score function to steer the sample generation towards desired properties. Such information is coined as guidance. For example, in text-to-image synthesis, text input is encoded as guidance to generate semantically aligned images. Proper guidance inputs are closely tied to the performance of diffusion models. A common observation is that strong guidance promotes a tight alignment to the task-specific information, while reducing the diversity of the generated samples. In this paper, we provide the first theoretical study towards understanding the influence of guidance on diffusion models in the context of Gaussian mixture models. Under mild conditions, we prove that incorporating diffusion guidance not only boosts classification confidence but also diminishes distribution diversity, leading to a reduction in the differential entropy of the output distribution. Our analysis covers the widely adopted sampling schemes including DDPM and DDIM, and leverages comparison inequalities for differential equations as well as the Fokker-Planck equation that characterizes the evolution of probability density function, which may be of independent theoretical interest. Copyright © 2024, The Authors. All rights reserved.

关键词： Diffusion

Image data augmentation for improving performance of deep learning-based model in pathological lung segmentation

学校读者我要写书评

暂无评论

Image data augmentation for improving performance of deep le...

Proceedings of the Digital Image Computing: Technqiues and Applications (DICTA)

作者： Md. Shariful Alam Dadong Wang Arcot Sowmya School of Computer Science and Engineering University of New South Wales Sydney Australia Quantitative Imaging Data61 CSIRO Sydney Australia

Accurate segmentation of lung fields from chest X-ray (CXR) images is very important for subsequent analysis of many pulmonary diseases. Deep Neural Networks (DNN)-based methods have achieved remarkable progress in many image related tasks. However, their performance depends highly on the distribution of training and test samples, and they perform well if both training and test samples are from the same distribution. For example, DNN-based lung segmentation methods perform well on segmentation of healthy lung or lung with mild disease, however their performance is poor on lungs with severe abnormalities. Pulmonary opacification, which blurs the lung boundary, is one of the main reasons. A solution to this problem is data augmentation to increase the pool of training images, however despite the great success of traditional data augmentation techniques for natural images, they are not very effective for medical images. To simulate CXR images with opacification and low contrast, we present a novel image data augmentation technique in this study. To generate an augmented image, we first generate a random area inside the lung and then blur the area with a gaussian filter. Then, low contrast is simulated by adjusting the contrast and brightness. To evaluate the utility of the proposed augmentation technique, we applied it to images with different pulmonary diseases such as tuberculosis, pneumoconiosis and covid-19 from three public datasets as well as a private dataset and compared its effect on segmentation performance with traditional data augmentation techniques. Results suggest that the proposed technique outperforms traditional data augmentation techniques for all datasets on lung segmentation, in terms of Dice Coefficient (DC) and Jaccard Index (JI). Extensive experiments on multiple datasets validate the effectiveness of the proposed data augmentation technique.

关键词： Training Image segmentation Pathology Pulmonary diseases Digital images Lung Indexes

Integrating a novel irrigation approximation method with a process-based remote sensing model to estimate multi-years'winter wheat yield over the North China Plain

学校读者我要写书评

暂无评论

Journal of Integrative Agriculture 2023年第9期22卷 2865-2881页

作者： ZHANG Sha YANG Shan-shan WANG Jing-wen WU Xi-fang Malak HENCHIRI Tehseen JAVED ZHANG Jia-hua BAI Yun Research Center for Space Information and Big Earth Data College of Computer Science and TechnologyQingdao UniversityQingdao 266071P.R.China Aerospace Information Research Institute Chinese Academy of SciencesBeijing 100094P.R.China School of Surveying and Land Information Engineering Henan Polytechnic UniversityJiaozuo 454000P.R.China Department of Environmental Sciences Kohat University of Science and TechnologyKohat 26000Pakistan

Accurate estimation of regional winter wheat yields is essential for understanding the food production status and ensuring national food ***,using the existing remote sensing-based crop yield models to accurately reproduce the inter-annual and spatial variations in winter wheat yields remains challenging due to the limited ability to acquire irrigation information in water-limited ***,we proposed a new approach to approximating irrigations of winter wheat over the North China Plain(NCP),where irrigation occurs extensively during the winter wheat growing *** approach used irrigation pattern parameters(IPPs)to define the irrigation frequency and ***,they were incorporated into a newly-developed process-based and remote sensing-driven crop yield model for winter wheat(PRYM–Wheat),to improve the regional estimates of winter wheat over the *** IPPs were determined using statistical yield data of reference years(2010–2015)over the *** findings showed that PRYM–Wheat with the optimal IPPs could improve the regional estimate of winter wheat yield,with an increase and decrease in the correlation coefficient(R)and root mean square error(RMSE)of 0.15(about 37%)and 0.90 t ha–1(about 41%),*** data in validation years(2001–2009 and 2016–2019)were used to validate PRYM–*** addition,our findings also showed R(RMSE)of 0.80(0.62 t ha–1)on a site level,0.61(0.91 t ha–1)for Hebei Province on a county level,0.73(0.97 t ha–1)for Henan Province on a county level,and 0.55(0.75 t ha–1)for Shandong Province on a city ***,PRYM–Wheat can offer a stable and robust approach to estimating regional winter wheat yield across multiple years,providing a scientific basis for ensuring regional food security.

关键词： approximating irrigations process-based model remote sensing winter wheat yield North China Plain

MedLogic-AQA: Enhancing Medical Question Answering with Abstractive Models Focusing on Logical Structures

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zafar, Aizan Mishra, Kshitij Ekbal, Asif Department of Computer Science and Engineering Indian Institute of Technology Patna India School of AI and Data Science Indian Institute of Technology Jodhpur India

In Medical question-answering (QA) tasks, the need for effective systems is pivotal in delivering accurate responses to intricate medical queries. However, existing approaches often struggle to grasp the intricate logical structures and relationships inherent in medical contexts, thus limiting their capacity to furnish precise and nuanced answers. In this work, we address this gap by proposing a novel Abstractive QA system MEDLOGIC-AQA that harnesses First Order Logic (FOL) based rules extracted from both context and questions to generate well-grounded answers. Through initial experimentation, we identified six pertinent first-order logical rules, which were then used to train a Logic-Understanding (LU) model capable of generating logical triples for a given context, question, and answer. These logic triples are then integrated into the training of MEDLOGIC-AQA, enabling effective and coherent reasoning during answer generation. This distinctive fusion of logical reasoning with abstractive QA equips our system to produce answers that are logically sound, relevant, and engaging. Evaluation with respect to both automated and human-based demonstrates the robustness of MEDLOGIC-AQA against strong baselines. Through empirical assessments and case studies, we validate the efficacy of MEDLOGIC-AQA in elevating the quality and comprehensiveness of answers in terms of reasoning as well as informativeness 1 © 2024, CC BY.

关键词：

EVDNET: Towards Explainable Multi Scale, Anchor Free Vehicle Detection Network in High Resolution Aerial Imagery

学校读者我要写书评

暂无评论

EVDNET: Towards Explainable Multi Scale, Anchor Free Vehicle...

IEEE International Symposium on Geoscience and Remote Sensing (IGARSS)

作者： Nandini Saini Shubham Gupta Chiranjoy Chattopadhyay Debasis Das Suman Kundu Computer Science & Engineering Indian Institute of Technology Jodhpur Rajasthan India School of Computing and Data Sciences FLAME University Pune Maharashtra India

ISBN: (数字)9798350360325

ISBN: (纸本)9798350360332

The rapid advancement in deep learning-based object detection methods has made them a prevalent choice for real-time applications. Families of object detectors, including one-stage detectors, two-stage detectors, and region-based CNN networks, offer superior performance in accurately detecting objects. Despite their high accuracy, the complex design and black-box functionality of these models are not directly transferable in aerial imagery. Also, raise questions among users regarding the transparency of the algorithm in locating objects. Consequently, to demystify the decision process of these models, there is a need for Explainable AI (XAI) tools. XAI enables an understanding of the significance of each pixel in an image, shedding light on the contributions that lead to the model’s final output. In this context, this work will present an efficient, explainable, multi-scale vehicle detection network from high resolution aerial imagery, named as EVDNet. The EVDNet model has trained with two publicly available aerial image benchmark dataset DOTA and VEDAI. To enhance interpretability, we leverage XAI method using GradCam. The experimental results not only showcase the effectiveness and performance of the EVDNet model but also provide valuable insights into the object detection process. This research contributes to bridging the gap between complex object detection models and user understanding, offering a more transparent and interpretable approach to high-resolution aerial imagery analysis.

关键词： Visualization Accuracy Image resolution Explainable AI Error analysis Vehicle detection Object detection

ProMamba: Prompt-Mamba for polyp segmentation

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Xie, Jianhao Liao, Ruofan Zhang, Ziang Yi, Sida Zhu, Yuesheng Luo, Guibo School of Electronic and Computer Engineering Peking University China Aberdeen Institute of Data Science and Artificial Intelligence South China Normal University China

Detecting polyps through colonoscopy is an important task in medical image segmentation, which provides significant assistance and reference value for clinical surgery. However, accurate segmentation of polyps is a challenging task due to two main reasons. Firstly, polyps exhibit various shapes and colors. Secondly, the boundaries between polyps and their normal surroundings are often unclear. Additionally, significant differences between different datasets lead to limited generalization capabilities of existing methods. To address these issues, we propose a segmentation model based on Prompt-Mamba, which incorporates the latest Vision-Mamba and prompt technologies. Compared to previous models trained on the same dataset, our model not only maintains high segmentation accuracy on the validation part of the same dataset but also demonstrates superior accuracy on unseen datasets, exhibiting excellent generalization capabilities. Notably, we are the first to apply the Vision-Mamba architecture to polyp segmentation and the first to utilize prompt technology in a polyp segmentation model. Our model efficiently accomplishes segmentation tasks, surpassing previous state-of-the-art methods by an average of 5% across six datasets. Furthermore, we have developed multiple versions of our model with scaled parameter counts, achieving better performance than previous models even with fewer parameters. Our code and trained weights will be released soon. Copyright © 2024, The Authors. All rights reserved.

关键词： Medical imaging