MEMLNaut Firmware: Multi-euclidean Sequencer with Reinforement Learning
Overview
A bank of 8 euclidean sequencers are controlled by a neural network, mapped from 3 axes of the joystick.
Euclidean Sequencing
Euclidean sequencing describes a process that generates events according to two numbers: N and K. N is the number of divisions within a time period. K is the number of events the algorithm tries to fit into these divisions. If N is divisible by K then a regularly spaced pattern is generated, otherwise the closest fit is chosen, creating varied rhythms.
A good analysis of euclidean rhythms can be found in Godfried Toussaint’s paper The Euclidean Algorithm Generates Traditional Musical Rhythms
Outputs
MIDI notes are output each time a euclidean generator triggers an new event.
Reinforcement Learning
When you indicate a reward, you are rewarding (negatively or positively) how a particular joystick position is mapped to the sequence it’s generating. As you indicate preferences for these settings, the DDPG algorithm will start to tune how the actor reacts to the joystick.