检索结果-内蒙古大学图书馆

1st Deep Breast Workshop on AI and Imaging for Diagnostic and Treatment Challenges in Breast Care, Deep-Breath 2024

作者： Hao, Degan Arefan, Dooman Zuley, Margarita L. Berg, Wendie A. Wu, Shandong Intelligent Systems Program University of Pittsburgh PittsburghPA United States Department of Radiology University of Pittsburgh PittsburghPA United States Department of Bioengineering University of Pittsburgh PittsburghPA United States Department of Biomedical Informatics University of Pittsburgh PittsburghPA United States

ISBN: (纸本)9783031777882

Adversarial data can lead to malfunction of deep learning applications. It is essential to develop deep learning models that are resilient to adversarial data while accurate on standard, clean data. In this study, we focus on building safe breast cancer diagnosis models against mammographic adversarial samples. We proposed a novel adversarially robust feature learning (ARFL) method to facilitate adversarial training using both standard data and adversarial data, where a feature correlation measure is incorporated as an objective function to encourage learning of robust features and restrain spurious features. To show the efficacy of ARFL for robust breast cancer diagnosis, we built and evaluated deep learning diagnosis models using two independent clinically collected breast imaging datasets, comprising a total of 9,548 mammogram images. We performed extensive experiments showing that the ARFL method outperformed several state-of-the-art methods. ARFL can serve as an effective method to enhance adversarial training, towards building safe breast cancer diagnosis against adversarial attacks in clinical settings. The code repository of this study is publicly available at GitHub: https://***/usernamesafeai/ARFL. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

Decarbonizing the Food System with Microbes and Carbon-Neutral Feedstocks

引用

Annual Review of Food Science and Technology 2025年第1期16卷 81-104页

作者： Mahmud, Niaz Taiwo, Kayode J. Usack, Joseph G. Sustainable Food Systems Engineering Group Department of Food Science and Technology University of Georgia AthensGA United States Food and Nutritional Sciences Program North Carolina A&T State University GreensboroNC United States New Materials Institute University of Georgia AthensGA United States Institute for Integrative and Precision Agriculture University of Georgia AthensGA United States

Harnessing CO2 and CO2-derived C1-C2 compounds for microbial food production can mitigate greenhouse gas emissions and boost sustainability within the food sector. These innovative technologies support carbon neutrality by generating nutrient-rich edible microbial biomass and biocompounds using autotrophic and heterotrophic microbes. However, qualifying microbial food viability and future impacts in the food sector remains challenging due to their diversity, technical complexity, socioeconomic forces, and incipient markets. This review provides an overview of microbial food systems and then delves into the technical interplay among feedstocks, microbes, carbon fixation platforms, bioreactor operations, and downstream processes. The review further explores developing markets for microbial food products, the industrial landscape, economic drivers, and emerging trends in next-generation food products. The analysis suggests a transformative shift in the food industry is underway, yet significant challenges persist, such as securing cost-effective feedstocks, improving downstream processing efficiency, and gaining consumer acceptance. These challenges require innovative solutions and collaborative efforts to ensure the future commercial success of microbial foods - doing so will create myriad opportunities to transform and decarbonize our food system. © 2025 by the author(s).

关键词： Feedstocks

来源：评论

学校读者我要写书评

暂无评论

Deep Learning for Real-Time PPE Usage Monitoring Using Wearable IMU Sensors

引用

IEEE Access 2025年 13卷 82492-82502页

作者： Guimarães, Pedro Carvalho da Fonseca De Cristo, Leonardo Braga Monteiro, Marcos Eduardo Pivaro Rebelatto, João Luiz Peron, Guilherme de Santi Rayel, Ohara Kerusauskas Moritz, Guilherme Luiz Federal University of Technology - Paraná (UTFPR) Curitiba Brazil Federal University of Technology - Paraná (UTFPR) Graduate Program in Electrical and Computer Engineering (CPGEI-CT) Curitiba Brazil Federal University of Technology - Paraná (UTFPR) Graduate Program in Energy Systems (PPGSE-CT) Curitiba Brazil

Ensuring the proper use of Personal Protective Equipment (PPE) is critical for safeguarding workers in high-risk environments, such as power distribution systems. This work presents an innovative approach to monitoring PPE usage through Deep Neural Networks (DNNs). Using data from inertial measurement units (IMUs) integrated into PPE items, we propose a solution to classify three usage states—namely carrying, still, and wearing—using raw accelerometer and gyroscope data. More specifically, thiswork assesses the effectiveness of three distinct network architectures—Convolutional Neural Networks (CNN), Bidirectional Long Short-Term Memory (BiLSTM), and CNN-BiLSTM—on a publicly released dataset on PPE usage provided by this study, while also comparing them against a traditional baseline Multi- Layer Perceptron (MLP) architecture. Our results demonstrate the superiority of BiLSTM in balancing high accuracy with computational efficiency, achieving above 98% accuracy regardless of the PPE type. Thiswork represents the first application of DNNs for PPE monitoring using IMU data, offering significant implications for enhancing safety compliance and operational monitoring in power distribution. © 2013 IEEE.

关键词： AI BiLSTM CNN DNN IoT PPE

来源：评论

学校读者我要写书评

暂无评论

Generative assistant for digital twin simulations 12

Generative assistant for digital twin simulations

引用

12th CIRP Global Web Conference, CIRPe 2024

作者： Boareto, Pedro Antonio De Freitas Rocha Loures, Eduardo Portela Santos, Eduardo Alves Deschamps, Fernando Industrial and Systems Engineering Graduate Program Pontifical Catholic University of Paraná 1155 Imaculada Conceição St. Paraná Curitiba80215-901 Brazil Federal University of Paraná 632 Prefeito Lothário Meissner Av. Paraná Curitiba80210-170 Brazil

One of the key emerging technologies in Industry 4.0 is the Digital Twin (DT). Although it promises increased efficiency, productivity, and innovation, its adoption faces challenges such as high investment costs and the need for workforce requalification. Generative Artificial Intelligence (GAI) emerges as a promising solution, offering capabilities to accelerate development processes and reduce costs. This study aims to leverage GAI to enhance the development of DT and support decision-making in industrial environments by proposing a Generative Assistant for Digital Twin Simulations (GADTS). This proposal generates operational models quickly, offers greater customization, and facilitates the creation of efficient scenario simulations in natural language. The proposal was tested with artificial data. As a result, the development of highly personalized DT simulations with Key Performance Indicators (KPIs) was entirely abstracted into natural language requests. © 2025 Elsevier B.V.. All rights reserved.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

BUFFER-X: Towards Zero-Shot Point Cloud Registration in Diverse Scenes

arXiv

引用

arXiv 2025年

作者： Seo, Minkyun Lim, Hyungtae Lee, Kanghee Carlone, Luca Park, Jaesik Interdisciplinary Program of AI Seoul National University Korea Republic of Laboratory for Information & Decision Systems Massachusetts Institute of Technology United States

Recent advances in deep learning-based point cloud registration have improved generalization, yet most methods still require retraining or manual parameter tuning for each new environment. In this paper, we identify three key factors limiting generalization: (a) reliance on environment-specific voxel size and search radius, (b) poor out-of-domain robustness of learning-based keypoint detectors, and (c) raw coordinate usage, which exacerbates scale discrepancies. To address these issues, we present a zero-shot registration pipeline called BUFFER-X by (a) adaptively determining voxel size/search radii, (b) using farthest point sampling to bypass learned detectors, and (c) leveraging patch-wise scale normalization for consistent coordinate bounds. In particular, we present a multi-scale patch-based descriptor generation and a hierarchical inlier search across scales to improve robustness in diverse scenes. We also propose a novel generalizability benchmark using 11 datasets that cover various indoor/outdoor scenarios and sensor modalities, demonstrating that BUFFER-X achieves substantial generalization without prior information or manual parameter tuning for the test datasets. Our code is available at https://***/MIT-SPARK/BUFFER-X. © 2025, CC BY-NC-ND.

关键词： Zero-shot learning

来源：评论

学校读者我要写书评

暂无评论

Cognitive AI framework: advances in the simulation of human thought

arXiv

引用

arXiv 2025年

作者： Salas-Guerra, Rommel AU University San Juan Puerto Rico Doctoral Program in Information Systems Management AGM University OrlandoFL United States

The Human Cognitive Simulation Framework constitutes a paradigmatic advance in integrating human cognitive capabilities in artificial intelligence systems, optimizing personalization and adaptability in human-AI interaction. This model merges contexts of short-term memory (conversation context), long-term memory (interaction context), advanced cognitive processing, and efficient knowledge management, guaranteeing contextual coherence during sessions and the storage of persistent information, which will enhance personalization and continuity in future AI interactions with the user. This framework emphasizes the relevance of a unified database that synchronizes these contexts, ensuring accessibility and operational efficiency, highlighting the incorporation of logical, creative, and analog processing modules like those of the left and right hemispheres of the human brain, which serve to perform structured tasks and complex contextual inferences. Therefore, in Cognitive AI, the dynamic updating of knowledge allows the integration of information in real-time, enriching the pre-trained knowledge of the model and ensuring its adaptability, being useful in particularly relevant applications such as education, human behavior and knowledge management, promoting the development of adaptive learning platforms, promoting tools for personalized psychological interventions and generating Assertive synthetic knowledge. Although Cognitive AI can process large volumes of data and improve the user experience, it faces challenges associated with scalability, mitigation of cognitive biases and compliance with ethical standards, therefore, this framework establishes a robust basis for future research, proposing topics such as continuous learning algorithms, sustainability and multimodal adaptability in information processing devices, so that it can consolidate itself as a transformative model in emerging disciplines. © 2025, CC BY-NC-ND.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Towards an Integration of Augmented Reality in Industrial Assembly Processes: Issues and Perspectives

引用

Procedia Computer Science 2025年 253卷 1063-1072页

作者： Bruno J. Souza Anderson L. Szejka Roberto Z. Freire Industrial and Systems Engineering Graduate Program Pontifical Catholic University of Parana (PUCPR) Curitiba 80215-901 Brazil Federal University of Technology – Parana (UTFPR) Curitiba 80050-540 Brazil

The evolution of manufacturing has led to Industry 5.0, where human creativity is integrated with advanced technology to achieve efficient and sustainable production. Given the high volume and diversity of assembly tasks, operator assistance systems have become essential tools in the industry, leveraging augmented reality (AR) and other new technologies. This literature review specifically examines 19 papers on AR applications within the context of industrial assembly tasks. The review focuses on identifying the types of AR technologies utilized, the methodologies for implementing AR systems in assembly, and the primary limitations and challenges encountered. The findings indicate a growing field with promising applications in assembly, training, and maintenance, predominantly using AR glasses. Furthermore, a gap is identified indicating the absence of a standardized framework for AR-based assembly assistance systems. Future research should aim at creating a foundational framework for AR system implementation, thereby accelerating the adoption of innovative technologies in the industry.

关键词： Industry 5.0 augmented reality assembly assistance

来源：评论

学校读者我要写书评

暂无评论

Survey of the Tsuzuraozaki Underwater Archaeological Site Using Hovering Type Autonomous Underwater Vehicle

Survey of the Tsuzuraozaki Underwater Archaeological Site Us...

引用

2025 IEEE Underwater Technology, UT 2025

作者： Miyakawa, Ryo Odasaki, Issei Yano, Kenichi Kumagai, Michio Sugimatsu, Harumi Inoue, Shogo Nishida, Yuya Kyushu Institute of Technology Dept. of Life Science and Systems Engineering Kitakyushu Japan Kyushu Institute of Technology Dept. of Human Inteligence Systems Kitakyushu Japan College of Letters Japanese History Program Ritsumeikan University Kyoto Japan Ritsumeikan University Research Organization of Science and Technology Kyoto Japan Institute of Industrial Science The University of Tokyo Meguro Japan IDEA Consultants Inc. Ocean Research Department Setagaya Japan

ISBN: (纸本)9798331534189

Many pottery have been discovered in the waters around Tsuzuraozaki, which is located on the northern shore of Lake Biwa in Shiga Prefecture. However the cause of the archaeological site has not yet been determined, because of the depth and high turbidity of the water, it is difficult for divers to investigate the area. Therefore, in December 2023, we conducted an observation survey of the lake bed by mounting a camera on an Autonomous Underwater Vehicle Tuna-Sand2. We created a mosaic image by enhancing the image with Malti-Scale Retinex and pasting it along the trajectory of the AUV. The survey results showed the location of the pottery-like objects and the surrounding terrain. However, it is difficult to determine from the image whether the object is pottery or not, and in which period it was made. In future research, we aim to delve deeper into the intricacies of the objects found in this survey. In addition, it is also essential to have a technology that can remove floating matter from images of water with high turbidity. © 2025 IEEE.

关键词： Autonomous underwater vehicles

来源：评论

学校读者我要写书评

暂无评论

Evaluation of Cuffless Blood Pressure Estimation Accuracy at Each Heartbeat During Exercise 6th

Evaluation of Cuffless Blood Pressure Estimation Accuracy a...

引用

6th International Conference on Biomedical and Health Informatics, ICBHI 2024

作者： Ono, Kouki Maeda, Yuka Yoshimura, Takumi Huang, Ming Tamura, Toshiyo Master’s Program in Intelligent and Mechanical Interaction Systems University of Tsukuba Tsukuba305-8577 Japan Institute of Systems and Information Engineering University of Tsukuba Tsukuba305-8577 Japan Medical and Welfare Engineering Tokyo Metropolitan College of Industrial Technology Tokyo116-8523 Japan School of Data Science Nagoya City University Nagoya467-8501 Japan Future Robotics Organization Waseda University Tokyo169-8050 Japan

ISBN: (纸本)9783031863226

Continuous blood pressure monitoring is crucial because of its dynamic nature, influenced by factors such as physical activity and mental stress. Although cuffless blood pressure estimation is a promising technique for continuous monitoring, its accuracy during drastic fluctuations, especially during exercise, remains insufficiently validated. In this study, we evaluated a simple Pulse Arrival Time based blood pressure estimation method and several improved methods for sharp fluctuations induced by exercise. Sixteen subjects in the experiment, with data from 13 valid for analysis. The proposed method achieved an average correlation of 0.527 in systolic blood pressure, suggesting that cuffless blood pressure estimation could be applied to estimate continuous blood pressure variability during exercise. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Blood pressure

来源：评论

学校读者我要写书评

暂无评论

Generative a bistant for digital twin simulations 1

Generative a bistant for digital twin simulations

引用

1st International Conference on Digital Sovereignty, ICDS 2024

作者： Boareto, Pedro Antonio De Freitas Rocha Loures, Eduardo Portela Santos, Eduardo Alves Deschamps, Fernando Industrial and Systems Engineering Graduate Program Pontifical Catholic University of Paraná 1155 Imaculada Conceição St. Curitiba Paraná80215-901 Brazil Federal University of Paraná 632 Prefeito Lothário Mei bner Av. Curitiba Paraná80210-170 Brazil

One of the key emerging technologies in Industry 4.0 is the Digital Twin (DT). Although it promises increased efficiency, productivity, and innovation, its adoption faces challenges such as high investment costs and the need for workforce requalification. Generative Artificial Intelligence (GAI) emerges as a promising solution, offering capabilities to accelerate development proce bes and reduce costs. This study aims to leverage GAI to enhance the development of DT and support decision-making in industrial environments by proposing a Generative A bistant for Digital Twin Simulations (GADTS). This proposal generates operational models quickly, offers greater customization, and facilitates the creation of efficient scenario simulations in natural language. The proposal was tested with artificial data. As a result, the development of highly personalized DT simulations with Key Performance Indicators (KPIs) was entirely abstracted into natural language requests. © 2025 Elsevier B.V.. All rights reserved.

关键词： Generative adversarial networks

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：