Learning to Create (Lrn2Cre8)

Lrn2Cre8 aims to understand the relationship between learning and creativity by means of practical engineering, theoretical study, and cognitive comparison.

The role of the SONY Music Team consists essentially in evaluating the impact of computational-creative outputs by creating a dedicated music label, and to develop novel methodologies for analytic and social evaluation of such artefacts in an attempt to bridge the gap between two classes of traditionally incompatible musical systems: 1) interactive musical systems, limited in their ability to generate stylistically consistent material, and 2) music composition systems, which are fundamentally not interactive.

One of the first activities consisted in drafting the report "Music Generation with Neural Networks", in order to explain how methods such as Recurrent Neural Networks or Long Short-Term Memory and their variants can be applied to music generation. Here are some examples related to our analysis on music generation techniques:

Example of generation with a LSTM network
Example of generation with a Bidirectional LSTM network
Example of generation with a Multi-dimensional LSTM network










Home >> Lrn2Cre8