Citation, DOI, disclosures and article data
At the time the article was created Andrew Murphy had no recorded disclosures.View Andrew Murphy's current disclosures
At the time the article was last revised Rohit Sharma had no recorded disclosures.View Rohit Sharma's current disclosures
Machine learning is a specific practical application of computer science and mathematics that allows computers to extrapolate information based on observed patterns without explicit programming. A defining characteristic of machine learning programs is the improved performance at tasks such as classification when more data, known as training data, is processed.
Machine learning is becoming an increasingly important tool in the medical profession for primary computer-aided diagnosis algorithms and decision support systems. Interest in the practical applications of machine learning, including applications for imaging, is high. The availability of large scale data sets, substantial advances in computing power, and new deep-learning algorithms have driven this area forward substantially since it began in the 1950s. It is highly likely that in the next decade various implementations of machine learning will have a profound impact on the way radiology is practiced. To many in the field, it appears inevitable that many of the tasks that are currently considered core to the practice of radiology (e.g. abnormality detection in images and classification of images) will be performed at least in part by these systems.
Although there are countless specific models and implementations of machine learning, the majority used in radiology fall into one of a relatively small number of fundamentally different underlying learning processes and models.
How machine learning processes are implemented is variable and determined in part by the type of problem they are designed to solve. Although much of the recent work in the field of image processing generally, and more specifically radiology, has focussed on convolutional neural networks, a type of neural network, a number of other models are useful in various circumstances. These include:
- linear regression
- logistic regression
- decision tree
- random forest
- support vector machines
- clustering models
History and etymology
Arthur Samuel first defined the term "machine learning" in 1959, a component in artificial intelligence where a computer learns from a set of data to improve its future performances 6.
- 1. Erickson BJ, Korfiatis P, Akkus Z, Kline TL. Machine Learning for Medical Imaging. Radiographics : a review publication of the Radiological Society of North America, Inc. 37 (2): 505-515. doi:10.1148/rg.2017160130 - Pubmed
- 2. LeCun Y, Bengio Y, Hinton G. Deep learning. Nature. 521 (7553): 436-44. doi:10.1038/nature14539 - Pubmed
- 3. Baştanlar Y, Ozuysal M. Introduction to machine learning. Methods in molecular biology (Clifton, N.J.). 1107: 105-28. doi:10.1007/978-1-62703-748-8_7 - Pubmed
- 4. Obermeyer Z, Emanuel EJ. Predicting the Future - Big Data, Machine Learning, and Clinical Medicine. The New England journal of medicine. 375 (13): 1216-9. doi:10.1056/NEJMp1606181 - Pubmed
- 6. European Society of Radiology (ESR). What the Radiologist Should Know About Artificial Intelligence – an ESR White Paper. Insights Imaging. 2019;10(1):44. doi:10.1186/s13244-019-0738-2 - Pubmed
- 5. Mitchell, T. (1997). Machine Learning. McGraw Hill. p. 2. ISBN 978-0-07-042807-2.