This allows the model to more easily repeat 1 or 2 bar phrases without having to store those sequences in its memory cell. For example, if the third bar of the melody is completely repeating the first bar, every label for that third bar will be the repeat-2-bars-ago label. Only when the melody isn’t repeating 1 or 2 bars ago do we make the label for that step be a specific melody event. If it’s not repeating the event from 2 bars ago, we check if it’s repeating the event from 1 bar ago, and if so, we set the label for that step to be repeat-1-bar-ago.
When creating labels for the training data, if the current event in the melody is repeating the same event from 2 bars ago, we set the label for that step to be repeat-2-bars-ago. This is where the Lookback RNN gets its name. The label to repeat the event from 1 bar ago and the label to repeat the event from 2 bars ago. In addition to feeding the model more input information, we also add two new custom labels. The only difference being the values are -1 and 1 instead of 0 and 1. These inputs are 5 values that can be thought of as a binary step clock. We also input the current position within the measure (as done previously by Daniel Johnson), allowing the model to more easily learn patterns associated with 4/4 time music. This allows the model to more easily recognize patterns associated with being in a repetitive or non-repetitive state. This signals if the last event was creating something new, or just repeating an already established melody. We also input whether the last event was repeating the event from 1 or 2 bars before it. This allows the model to more easily recognize patterns that occur across 1 and 2 bars, such as mirrored or contrasting melodies. In addition to inputting the previous event, we also input the events from 1 and 2 bars ago. In Lookback RNN, we add the following additional information to the input vector: The possible events were note-off (turn off any currently playing note), no event (if a note is playing, continue sustaining it, otherwise continue silence), and a note-on event for each pitch (which also turns off any other note that might be playing). In our introductory model, Basic RNN, the input to the model was a one-hot vector of the previous event, and the label was the target next event. The type of RNN cell used in this model is an LSTM. The custom labels make it easier for the model to repeat sequences of notes without having to store them in the RNN’s cell state. They also help the model recognize patterns related to where in the measure an event occurs. The custom inputs allow the model to more easily recognize patterns that occur across 1 and 2 bars. Lookback RNN introduces custom inputs and labels. We hope you’ll join us in exploring how they might produce better songs and stories. Without long-term structure, the content produced by recurrent neural networks (RNNs) often seems wandering and random.īut what if these RNN models could recognize and reproduce longer-term structure? Could they produce content that feels more meaningful – more human? Today we’re open-sourcing two new Magenta models, Lookback RNN and Attention RNN, both of which aim to improve RNNs’ ability to learn longer-term structures. Likewise, they can produce a screenplay with grammatically correct sentences, but not one with a compelling plot line. Basic machine learning systems can generate a short melody that stays in key, but they have trouble generating a longer melody that follows a chord progression, or follows a multi-bar song structure of verses and choruses.
Long-term structure comes very naturally to people, but it’s very hard for machines.
One of the difficult problems in using machine learning to generate sequences, such as melodies, is creating long-term structure.
0 kommentar(er)
