检索结果-内蒙古大学图书馆

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

学校读者我要写书评

暂无评论

Improving VTE Identification through Language Models from Ra...

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Jamie Deng Yusen Wu Yelena Yesha Phuong Nguyen Department of Computer Science University of Miami Miami FL USA Frost Institute for Data Science & Computing University of Miami Miami FL USA Department of Radiology University of Miami Miami FL USA

ISBN: (数字)9798350386226

ISBN: (纸本)9798350386233

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.

关键词： Deep learning Training Accuracy Computational modeling Biological system modeling Computer architecture Feature extraction Complexity theory Thrombosis Biomedical imaging

STRONG BASELINE AND BAG OF TRICKS FOR COVID-19 DETECTION OF CT SCANS

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Hsu, Chih-Chung Jian, Chih-Yu Lee, Chia-Ming Tsai, Chi-Han Dai, Sheng-Chieh Institute of Data Science Department of Statistics Center of Data Science Miin Wu School of Computing National Cheng Kung University No.1 University Rd Tainan City Taiwan

This paper investigates the application of deep learning models for lung Computed Tomography (CT) image analysis. Traditional deep learning frameworks encounter compatibility issues due to variations in slice numbers and resolutions in CT images, which stem from the use of different machines. Commonly, individual slices are predicted and subsequently merged to obtain the final result;however, this approach lacks slice-wise feature learning and consequently results in decreased performance. We propose a novel slice selection method for each CT dataset to address this limitation, effectively filtering out uncertain slices and enhancing the model’s performance. Furthermore, we introduce a spatial-slice feature learning (SSFL) technique[1] that employs a conventional and efficient backbone model for slice feature training, followed by extracting one-dimensional data from the trained model for COVID and non-COVID classification using a dedicated classification model. Leveraging these experimental steps, we integrate one-dimensional features with multiple slices for channel merging and employ a 2D convolutional neural network (CNN) model for classification. In addition to the aforementioned methods, we explore various high-performance classification models, ultimately achieving promising results. © 2023, CC BY.

关键词： COVID-19

An Analysis of the Efficacy of Rough Set Models Within the Framework of Binary Information Tables 11th

学校读者我要写书评

暂无评论

An Analysis of the Efficacy of Rough Set Models Within th...

11th International Joint Conference on Rough Sets, IJCRS 2025

作者： Kong, Qingzhao Yan, Conghao Department of Science Jimei University Xiamen361021 China Digital Fujian Big Data Modeling and Intelligent Computing Institute Jimei University Xiamen361021 China

ISBN: (纸本)9783031927461

Rough set (RS) theory is one of the most effective data mining tools. So far, researchers have proposed dozens or even hundreds of RS models to cope with various data challenges. However, scholars have not systematically compared and investigated the performance of these models in analyzing data. To address this issue, we select 12 most representative RS models derived from binary information tables and conduct comparative studies on them in terms of computational efficiency, approximation accuracy, classification accuracy, data dimensionality, and comprehensive performance. Through detailed experimental results, we reasonably evaluate the advantages of these models and offer reliable references for selecting suitable models to complete specific learning tasks. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Taxonomies

RIDE: Enhancing Large Language Model Alignment through Restyled In-Context Learning Demonstration Exemplars

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Hua, Yuncheng Qu, Lizhen Li, Zhuang Xue, Hao Salim, Flora D. Haffari, Gholamreza Department of Data Science & AI Monash University Australia School of Computer Science Engineering University of New South Wales Australia School of Computing Technologies Royal Melbourne Institute of Technology Australia

Alignment tuning is crucial for ensuring large language models (LLMs) behave ethically and helpfully. Current alignment approaches require high-quality annotations and significant training resources. This paper proposes a low-cost, tuning-free method using in-context learning (ICL) to enhance LLM alignment. Through an analysis of high-quality ICL demos, we identified style as a key factor influencing LLM alignment capabilities and explicitly restyled ICL exemplars based on this stylistic framework. Additionally, we combined the restyled demos to achieve a balance between the two conflicting aspects of LLM alignment—factuality and safety. We packaged the restyled examples as prompts to trigger few-shot learning, improving LLM alignment. Compared to the best baseline approach, with an average score of 5.00 as the maximum, our method achieves a maximum 0.10 increase on the Alpaca-eval task (from 4.50 → 4.60), a 0.22 enhancement on the just-eval-instruct benchmark (from 4.34 → 4.56), and a maximum improvement of 0.32 (from 3.53 → 3.85) on the MT-Bench dataset. We release the code and data at https://***/AnonymousCode-Computerscience/RIDE. © 2025, CC BY.

关键词： Contrastive Learning

Internet of Thing (IoT) and data Analytics with Challenges and Future Applications 1

学校读者我要写书评

暂无评论

Internet of Thing (IoT) and Data Analytics with Challenges a...

1st International Conference on Computer science and Emerging Technologies, CSET 2023

作者： Aljarrah, Mohammmadnour Mosbah Zawaideh, Farah H. Magableh, Murad Al Wahshat, Hassan Mohamed, Rajina R. Archana, V.K. Irbid National University Faculty of Business and Financial Science MIS Bi and Data Analysis Department Irbid Jordan Irbid National University Faculty of Financial Sciences and Business Department of Bi and Data Analysi Irbid Jordan Ajloun National University Faculty of Business Department of Management Information System Ajloun Jordan College of Computing Dan Informatics Universiti Tenaga Nasional Malaysia Jain Deemed to Be University Bangalore India

ISBN: (纸本)9798350341737

The Internet of Things (IoT) has emerged as a transformative technology, enabling seamless connectivity and communication between smart devices and the digital world. IoT's integration with data analytics has opened up a myriad of possibilities for extracting valuable insights from the vast volumes of data generated by connected devices. This research paper explores the convergence of IoT and data analytics, shedding light on the challenges encountered during their integration and envisioning future applications that can revolutionize various domains. The paper begins with a comprehensive review of the IoT ecosystem and its significance in shaping the modern era of interconnected devices. It highlights the diverse range of IoT applications across industries, including smart cities, healthcare, agriculture, manufacturing, transportation, and more. Moreover, it delves into the underlying technologies and communication protocols that enable IoT devices to gather, transmit, and process data efficiently. As data generated by IoT devices continues to escalate exponentially, effective data analytics becomes paramount to extract meaningful insights and support informed decision-making. The research delves into various data analytics techniques, such as machine learning, artificial intelligence, and big data analytics, illustrating their relevance in harnessing the potential of IoT-generated data. However, amidst the promises, several challenges impede the seamless amalgamation of IoT and data analytics. Issues concerning data privacy, security, interoperability, scalability, and data quality are thoroughly examined. The research emphasizes the need for robust frameworks and standards to address these challenges and ensure the secure and efficient functioning of IoT ecosystems. In the final section, this paper envisions the future applications of IoT and data analytics. It envisions intelligent and autonomous systems that adapt to dynamic environments, leading to advancement

关键词： Internet of things

Multi-Agent System Based data Mining Technique for Supplier Selection

学校读者我要写书评

暂无评论

Multi-Agent System Based Data Mining Technique for Supplier ...

2023 International Conference on Power Energy, Environment and Intelligent Control, PEEIC 2023

作者： Dhatterwal, Jagjit Singh Kaswan, Kuldeep Singh Singh, Jay Naruka, Mahaveer Singh Govardhan, D. Artificial Intelligence and Data Science Koneru Lakshmaiah Education Foundation AP Guntur India School of Computing Science and Engineering Galgotias University UP Greater Noida India G L Bajaj Institute of Technology and Management Department of Electrical and Electronics Engineering Greater Noida India Institute of Aeronautical Engineering Department of Aeronautical Engineering Hyderabad India

ISBN: (纸本)9798350357769

Different target knowledge representations, which correspond to different hypothesis spaces, are suitable for learning various types of target functions. A distinct learning algorithm, capitalizing on an underlying structure to guide the exploration of the hypothesis space, leverages each of these hypothesis representations. Consequently, making decisions on these matters entails navigating an expansive landscape of alternative approaches to determine the one that best aligns with the defined learning problem. To select an appropriate learning algorithm that performs optimally for the given problem and target function, it is essential to analyze the interplay between the size and completeness of the hypothesis space, the availability of training examples, the learner's prior knowledge, and the confidence in a hypothesis's ability to generalize correctly to unseen data points. This challenge often arises in the context of poorly integrated systems. The objective of this paper is to introduce a specific approach to addressing these challenges, involving the use of enhanced elicitation techniques and tools, improved development paradigms, knowledge modeling languages, ontologies, and advanced methods for system maintenance. However, in the past decade, an alternative reasoning paradigm and computational problem-solving approach have garnered significant attention, involving the use of Multi-Agent System (MAS) based data Mining (DM) techniques. This paper proposes the integration of MAS and DM for supplier selection, with a focus on partitioning-based algorithms such as K-means, Farthest First, Expectation Maximization, and Filtered Clusterer. These algorithms are compared based on factors such as dataset size, the number of clusters, and the time required for cluster formation. Finally, within this integrated model, MAS-DM is exemplified through an illustrative result-matching process. © 2023 IEEE.

关键词： Cluster data Mining JADE Multi-Agent Ontologies

BRAINOOD: OUT-OF-DISTRIBUTION GENERALIZABLE BRAIN NETWORK ANALYSIS

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Xu, Jiaxing Chen, Yongqiang Dong, Xia Lan, Mengcheng Huang, Tiancheng Bian, Qingtian Cheng, James Ke, Yiping College of Computing and Data Science Nanyang Technological University Singapore Department of Computer Science and Engineering The Chinese University of Hong Kong Hong Kong S-Lab Nanyang Technological University Singapore

In neuroscience, identifying distinct patterns linked to neurological disorders, such as Alzheimer’s and Autism, is critical for early diagnosis and effective intervention. Graph Neural Networks (GNNs) have shown promising in analyzing brain networks, but there are two major challenges in using GNNs: (1) distribution shifts in multi-site brain network data, leading to poor Out-of-Distribution (OOD) generalization, and (2) limited interpretability in identifying key brain regions critical to neurological disorders. Existing graph OOD methods, while effective in other domains, struggle with the unique characteristics of brain networks. To bridge these gaps, we introduce BrainOOD, a novel framework tailored for brain networks that enhances GNNs’ OOD generalization and interpretability. BrainOOD framework consists of a feature selector and a structure extractor, which incorporates various auxiliary losses including an improved Graph Information Bottleneck (GIB) objective to recover causal subgraphs. By aligning structure selection across brain networks and filtering noisy features, BrainOOD offers reliable interpretations of critical brain regions. Our approach outperforms 16 existing methods and improves generalization to OOD subjects by up to 8.5%. Case studies highlight the scientific validity of the patterns extracted, which aligns with the findings in known neuroscience literature. We also propose the first OOD brain network benchmark, which provides a foundation for future research in this field. Our code is available at https://***/AngusMonroe/BrainOOD. © 2025, CC BY-NC-SA.

关键词： Graph neural networks

Alzheimer's Disease Detection using Machine Learning: A Review 3

学校读者我要写书评

暂无评论

Alzheimer's Disease Detection using Machine Learning: A Revi...

3rd International Conference on Advances in computing, Communication Control and Networking, ICAC3N 2021

作者： Dahiya, Savita Vijayalakshmi, S. Sabharwal, Munish Galgotias University School of Computing Science and Engineering Greater Noida India Christ University Department of Data Science Lavasa India Galgotias University School of Computing Science Greater Noida India

ISBN: (纸本)9781665438117

Alzheimer's is a progressive brain disorder which is an untreatable, and inoperable and mostly affect the elderly people. There is a new case of Alzheimer's disease being discovered globally in every four seconds. The outcome is fatal, as it results in death. Timely identification of Alzheimer's disease can be beneficial for us to get necessary care and possibly even avert brain tissue damage by the time. Effective automated techniques are required for detecting Alzheimer's disease at very early stage. Researchers use a variety of novel approaches to classify Alzheimer's disease. machine learning, an AI branch use probabilistic technique that allow system to acquire knowledge from huge amount of data. In this paper we represent a analysis report of the work which is done by researcher in this field. Research has achieved quite promising prediction accuracies however they were evaluated the the non-existent datasets from various imaging modalities which makes it difficult to make the fair comparison with the other methods comparison among them. In this paper, we conducted a study on the effectiveness of using human brain MRI scans to detect Alzheimer's disease and ended with a future discussion of Alzheimer's research trends. © 2021 IEEE.

关键词： Support vector machines

"Owls are wise and foxes are unfaithful": Uncovering animal stereotypes in vision-language models

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Aman, Tabinda Nadeem, Mohammad Sohail, Shahab Saquib Anas, Mohammad Cambria, Erik Department of Computer Science Aligarh Muslim University UP Aligarh202002 India School of Computing Science and Engineering VIT Bhopal University MP 466114 India Department of Computer Science and Engineering School of Engineering Sciences and Technology Jamia Hamdard New Delhi110062 India College of Computing and Data Science Nanyang Technological University Singapore

Animal stereotypes are deeply embedded in human culture and language. They often shape our perceptions and expectations of various species. Our study investigates how animal stereotypes manifest in vision-language models during the task of image generation. Through targeted prompts, we explore whether DALL-E perpetuates stereotypical representations of animals, such as "owls as wise," "foxes as unfaithful," etc. Our findings reveal significant stereotyped instances where the model consistently generates images aligned with cultural biases. The current work is the first of its kind to examine animal stereotyping in vision-language models systematically and to highlight a critical yet underexplored dimension of bias in AI-generated visual content. Copyright © 2025, The Authors. All rights reserved.

关键词： Stereo image processing