[Published paper] SymCHM—An Unsupervised Approach for Pattern Discovery in Symbolic Music with a Compositional Hierarchical Model

We are thrilled that a new paper on the Compositional Hierarchical Model has just been published.

The paper describes a modified version of the model, SymCHM, capable of working with symbolic music representations. The paper describes the work done during the last year of the model development for the pattern discovery task.

The paper publicly available here: http://www.mdpi.com/2076-3417/7/11/1135

We are also thrilled to announce further research on this topic in a connection with the Slovene folk song datasets. The SymCHM is proving to be a great tool for music analysis and automatic melodic classification. More about this topic will be made available in the next few months.

ABSTRACT

This paper presents a compositional hierarchical model for pattern discovery in symbolic music. The model can be regarded as a deep architecture with a transparent structure. It can learn a set of repeated patterns within individual works or larger corpora in an unsupervised manner, relying on statistics of pattern occurrences, and robustly infer the learned patterns in new, unknown works. A learned model contains representations of patterns on different layers, from the simple short structures on lower layers to the longer and more complex music structures on higher layers. A pattern selection procedure can be used to extract the most frequent patterns from the model. We evaluate the model on the publicly available JKU Patterns Datasets and compare the results to other approaches.