Deep learning

Digital Health and AI

Definition. Deep learning is a branch of machine learning in which neural networks with many layers learn representations of data directly from raw inputs such as images, signals or text.

Explanation

Each layer of a deep network transforms the output of the previous one, so that the model builds increasingly abstract features without hand-crafted variables. LeCun, Bengio and Hinton (2015) describe how backpropagation, large datasets and graphics processors made this approach dominant in image, speech and language tasks.

In health care, deep learning underlies image classification in radiology, pathology, dermatology and ophthalmology, prediction from electronic records and the large language models now used for documentation. Esteva et al. (2019) note that performance in retrospective datasets often falls in new settings, so external validation, calibration and monitoring of drift are as important as accuracy.

Related concepts

Machine learning Artificial intelligence (AI) Large language models (LLMs) Algorithmic bias

Read on The Health Topics

References

Esteva, A., Robicquet, A., Ramsundar, B., Kuleshov, V., DePristo, M., Chou, K., Cui, C., Corrado, G., Thrun, S., & Dean, J. (2019). A guide to deep learning in healthcare. Nature Medicine, 25(1), 24–29.

LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436–444.

Cite this entry: Kurutkan, M. N. (2026). Deep learning. In The Health Topics glossary. https://healthtopic.org/glossary/deep-learning/

Articles that discuss this concept

36 articles on The Health Topics use this concept; the 12 that use it most are listed. The list updates as new articles are published.

Explore the knowledge graph →

← Back to the glossary