Finnish perspective on using synthetic health data to protect privacy: the PRIVASA project DOI

Tinja Pitkämäki,

Tapio Pahikkala, Ileana Montoya Perez

и другие.

Applied Computing and Intelligence, Год журнала: 2024, Номер 4(2), С. 138 - 163

Опубликована: Янв. 1, 2024

<p>The use of synthetic data could facilitate data-driven innovation across industries and applications. Synthetic can be generated using a range methods, from statistical modeling to machine learning generative AI, resulting in datasets different formats utility. In the health sector, is often motivated by privacy concerns. As AI becoming an everyday tool, there need for practice-oriented insights into prospects limitations data, especially sensitive domains. We present interdisciplinary outlook on topic, focusing on, but not limited to, Finnish regulatory context. First, we emphasize working definitions avoid misplaced assumptions. Second, consider cases viewing it as helpful tool experimentation, decision-making, building literacy. Yet complementary uses should diminish continued efforts collect share high-quality real-world data. Third, discuss how privacy-preserving fall existing protection frameworks. Neither process generation nor are automatically exempt obligations concerning personal Finally, explore future research directions generating conclude discussing potential developments at societal level.</p>

Язык: Английский

Computer-aided detection systems based on ballistocardiography signals: A review DOI Creative Commons
Dalibor Cimr, Damián Bušovský, Hamido Fujita

и другие.

Engineering Applications of Artificial Intelligence, Год журнала: 2025, Номер 151, С. 110669 - 110669

Опубликована: Апрель 12, 2025

Язык: Английский

Процитировано

0
A novel CTGAN-ENN hybrid approach to enhance the performance and interpretability of machine learning black-box models in intrusion detection and IoT DOI
Houssam Zouhri, Ali Idri

Future Generation Computer Systems, Год журнала: 2025, Номер unknown, С. 107882 - 107882

Опубликована: Май 1, 2025

Язык: Английский

Процитировано

0
Adaptive Mask-Based Interpretable Convolutional Neural Network (AMI-CNN) for Modulation Format Identification DOI Creative Commons

Xiyue Zhu,

Yu Cheng,

Jiafeng He

и другие.

Applied Sciences, Год журнала: 2024, Номер 14(14), С. 6302 - 6302

Опубликована: Июль 19, 2024

Recently, various deep learning methods have been applied to Modulation Format Identification (MFI). The interpretability of models is important. However, this challenged due the black-box nature learning. To deal with difficulty, we propose an Adaptive Mask-Based Interpretable Convolutional Neural Network (AMI-CNN) that utilizes a mask structure for feature selection during neural network training and feeds selected features into classifier decision making. During training, masks are updated dynamically parameters optimize selection. extracted serves as interpretable weights, each weight corresponding feature, reflecting contribution model’s decision. We validate model on two datasets—Power Spectral Density (PSD) constellation phase histogram—and compare it three classical methods: Gradient-Weighted Class Activation Mapping (Grad-CAM), Local Model-Agnostic Explanations (LIME), Shapley Additive exPlanations (SHAP). MSE values follows: AMI-CNN achieves lowest 0.0246, followed by SHAP 0.0547, LIME 0.0775, Grad-CAM 0.1995. Additionally, highest PG-Acc 1, whether PSD or histogram. Experimental results demonstrate outperforms compared in both qualitative quantitative analyses.

Язык: Английский

Процитировано

1
Curvature index of image samples used to evaluate the interpretability informativeness DOI
Zhuo Zhang, Shuai Xiao, Meng Xi

и другие.

Engineering Applications of Artificial Intelligence, Год журнала: 2024, Номер 137, С. 109044 - 109044

Опубликована: Авг. 8, 2024

Язык: Английский

Процитировано

1
Automating Athletic Excellence Through Intelligent Process Automation and Sports Analytics DOI

Mohandass Lingappan,

Swaminathan Sethu,

T. Parasuraman

и другие.

Advances in computational intelligence and robotics book series, Год журнала: 2024, Номер unknown, С. 179 - 200

Опубликована: Сен. 20, 2024

This study examines the profound influence of artificial intelligence (AI) on sports industry, including its effects games, training methods, fan involvement, and player well-being. text explores how is transforming several aspects industry by analysing current trends future predictions. AI-powered intelligent referees are being developed to enhance fairness accuracy refereeing, while personalised experiences created increase spectator engagement. Furthermore, implementation health aid virtual reality environments expected performance raise safety standards. The integration technology athleticism in has potential revolutionise field AI, creating a mutually beneficial connection between innovation human accomplishment. will ultimately improve whole experience for everyone involved.

Язык: Английский

Процитировано

0
Finnish perspective on using synthetic health data to protect privacy: the PRIVASA project DOI

Tinja Pitkämäki,

Tapio Pahikkala, Ileana Montoya Perez

и другие.

Applied Computing and Intelligence, Год журнала: 2024, Номер 4(2), С. 138 - 163

Опубликована: Янв. 1, 2024

<p>The use of synthetic data could facilitate data-driven innovation across industries and applications. Synthetic can be generated using a range methods, from statistical modeling to machine learning generative AI, resulting in datasets different formats utility. In the health sector, is often motivated by privacy concerns. As AI becoming an everyday tool, there need for practice-oriented insights into prospects limitations data, especially sensitive domains. We present interdisciplinary outlook on topic, focusing on, but not limited to, Finnish regulatory context. First, we emphasize working definitions avoid misplaced assumptions. Second, consider cases viewing it as helpful tool experimentation, decision-making, building literacy. Yet complementary uses should diminish continued efforts collect share high-quality real-world data. Third, discuss how privacy-preserving fall existing protection frameworks. Neither process generation nor are automatically exempt obligations concerning personal Finally, explore future research directions generating conclude discussing potential developments at societal level.</p>

Язык: Английский

Процитировано

0