Self-supervised learning approach can test 20 million cells or more
- Researchers at the Technical University of Munich and Helmholtz Munich tested self-supervised learning for analyzing 20 million cells or more.
- The study shows that self-supervised learning enhances performance in transfer tasks, particularly with smaller datasets informed by larger ones.
- Masked learning is found to be more effective for applications involving large single-cell data sets compared to contrastive learning.
- The researchers aim to use machine learning to reinterpret existing datasets and derive insights about cell structure changes due to conditions like smoking and lung cancer.
5 Articles
5 Articles
Artificial intelligence in biomedicine: A key to analyzing millions of individual cells
Our bodies are made up of around 75 billion cells. But what function does each individual cell perform and how greatly do a healthy person's cells differ from those of someone with a disease? To draw conclusions, enormous quantities of data must be analyzed and interpreted. For this purpose, machine learning methods are applied. Researchers have now tested self-supervised learning as a promising approach for testing 20 million cells or more.
Self-supervised learning approach can test 20 million cells or more
Our bodies are made up of around 75 billion cells. But what function does each individual cell perform and how greatly do a healthy person's cells differ from those of someone with a disease? To draw conclusions, enormous quantities of data must be analyzed and interpreted.
Coverage Details
Bias Distribution
- 100% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
