Improving Performance of Self-Organising Maps with Distance Metric Learning Method
This addresses accuracy issues in SOM for pattern recognition users, but it is incremental as it adapts an existing metric learning method to a specific neural network architecture.
The paper tackled the problem of poor accuracy in Self-Organising Maps (SOM) by changing the distance metric from Euclidean to a learned Mahalanobis matrix using Large Margin Nearest Neighbour, resulting in improved performance on classification tasks such as digit, letter, and face recognition.
Self-Organising Maps (SOM) are Artificial Neural Networks used in Pattern Recognition tasks. Their major advantage over other architectures is human readability of a model. However, they often gain poorer accuracy. Mostly used metric in SOM is the Euclidean distance, which is not the best approach to some problems. In this paper, we study an impact of the metric change on the SOM's performance in classification problems. In order to change the metric of the SOM we applied a distance metric learning method, so-called 'Large Margin Nearest Neighbour'. It computes the Mahalanobis matrix, which assures small distance between nearest neighbour points from the same class and separation of points belonging to different classes by large margin. Results are presented on several real data sets, containing for example recognition of written digits, spoken letters or faces.