Self-Organizing Maps (SOM) are an unsupervised learning technique used in the field of artificial intelligence and machine learning.

The SOM algorithm is based on the idea that the input data are organized into a two-dimensional topological map, in which nearby points in the map represent similar patterns in the input data. During training of the SOM model, the data are presented to the map, and the weights of the nodes are adjusted according to the patterns in the input data.

Self-organizing maps are useful for multidimensional data visualization and pattern detection in large data sets. They are also used in dimensionality reduction, which is the process of representing complex data in a lower dimensional space, facilitating data analysis and visualization.

Self-organizing maps differ from other artificial neural networks because they apply competitive learning as opposed to error correction learning. SOM uses a neighborhood function to preserve the topological properties of the input space.

Self-organizing maps have found applications in a variety of areas, including image and signal processing, document classification, gene identification, sensor data analysis, and exploration of large datasets.

Reference: Teuvo Kohonen. Self-Organized Formation of Topologically Correct Feature Maps..