Conditional neural sequence learners for generating drums’ rhythms

Abstract

Machine learning has shown a successful component of methods for automatic music composition. Considering music as a sequence of events with multiple complex dependencies on various levels of a composition, the long short-term memory-based (LSTM) architectures have been proven to be very efficient in learning and reproducing musical styles. The “rampant force” of these architectures, however, makes them hardly useful for tasks that incorporate human input or generally constraints. Such an example is the generation of drums’ rhythms under a given metric structure (potentially combining different time signatures), with a given instrumentation (e.g. bass and guitar notes). This paper presents a solution that harnesses the LSTM sequence learner with a feed-forward (FF) part which is called the “Conditional Layer”. The LSTM and the FF layers influence (are merged into) a single layer making the final decision about the next drums’ event, given previous events (LSTM layer) and current constraints (FF layer). The resulting architecture is called the conditional neural sequence learner (CNSL). Results on drums’ rhythm sequences are presented indicating that the CNSL architecture is effective in producing drums’ sequences that resemble a learnt style, while at the same time conform to given constraints; impressively, the CNSL is able to compose drums’ rhythms in time signatures it has not encountered during training (e.g. 17/16), which resemble the characteristics of the rhythms in the original data.