Efficient Deep Learning-Based Automated Diagnosis from Echocardiography with Contrastive Self-Supervised Learning

Information Science, Computer Science, Artificial Intelligence, Medicine, Medical Imaging, Cardiovascular System,
deep learning   self-supervised learning   contrastive learning   echocardiography   automated diagnosis   left ventricular hypertrophy   aortic stenosis  

Breakthrough in Automated Echocardiogram Diagnosis via Deep Learning: A Comparative Study of Self-Supervised Learning Methods Research Background With the rapid development of artificial intelligence and machine learning technologies, their role in medical imaging diagnosis is becoming increasingly significant. In particular, Self-Supervised Learni...

Graph Neural Networks with Multiple Prior Knowledge for Multi-omics Data Analysis

Graph Neural Networks with Multiple Prior Knowledge for Multi-omics Data Analysis

Information Science, Computer Science, Artificial Intelligence, Medicine, Environmental and Public Health, Oncology, Life Sciences, Cell Biology, Genetics,
graph neural networks   multi-omics data   contrastive learning   biological knowledge   cancer molecular classification  

Multiple Prior Knowledge Graph Neural Network in Multi-Omics Data Analysis Background Introduction Precision medicine is an important field for the future of healthcare as it provides personalized treatment plans for patients, improving treatment outcomes and reducing costs. For instance, due to the complex clinical, pathological, and molecular cha...

Mitigating Social Biases of Pre-trained Language Models via Contrastive Self-Debiasing with Double Data Augmentation

Information Science, Computer Science, Artificial Intelligence,
pre-trained language models   social bias   contrastive learning   prompt learning   data augmentation  

Introduction: Currently, pre-trained language models (PLMs) are widely applied in the field of natural language processing, but they have the problem of inheriting and amplifying social biases present in the training corpora. Social biases may lead to unpredictable risks in real-world applications of PLMs, such as automatic job screening systems te...