A Predictive Model for Music Based on Learned Interval Representations
This addresses a specific bottleneck in music AI for tasks like melody prediction and structure learning, though it is incremental relative to existing neural methods.
The paper tackled the problem of connectionist sequence models failing to generalize over relative musical concepts and capture repetition, by introducing the recurrent gated autoencoder (RGAE) that learns interval representations. The result showed that RGAE improves state-of-the-art in predicting monophonic melodies and greatly outperforms absolute pitch models on learning copy-and-shift operations.
Connectionist sequence models (e.g., RNNs) applied to musical sequences suffer from two known problems: First, they have strictly "absolute pitch perception". Therefore, they fail to generalize over musical concepts which are commonly perceived in terms of relative distances between pitches (e.g., melodies, scale types, modes, cadences, or chord types). Second, they fall short of capturing the concepts of repetition and musical form. In this paper we introduce the recurrent gated autoencoder (RGAE), a recurrent neural network which learns and operates on interval representations of musical sequences. The relative pitch modeling increases generalization and reduces sparsity in the input data. Furthermore, it can learn sequences of copy-and-shift operations (i.e. chromatically transposed copies of musical fragments)---a promising capability for learning musical repetition structure. We show that the RGAE improves the state of the art for general connectionist sequence models in learning to predict monophonic melodies, and that ensembles of relative and absolute music processing models improve the results appreciably. Furthermore, we show that the relative pitch processing of the RGAE naturally facilitates the learning and the generation of sequences of copy-and-shift operations, wherefore the RGAE greatly outperforms a common absolute pitch recurrent neural network on this task.