Predicting Reinforcement of Pitch Sequences via LSTM and TD

We examine the use of a recurrent neural network called Long Short-Term Memory (LSTM) with a prediction algorithm called temporal difference (TD) to predict the outcome of a music pitch sequence, while the sequence is being played. This is part of a larger system that will use this prediction in order to choose pitches to play. We describe our previous results using the LSTM network for musical tasks and then show its ability to predict a positive or negative outcome for a short musical task of chromatic lead-in to a chord tone. Then we describe its ability to predict positive outcomes when certain chord tones are played on the last beat of each bar of ii-V-I

[1]  Geoffrey E. Hinton,et al.  Learning internal representations by error propagation , 1986 .

[2]  Douglas H. Keefe,et al.  The Representation of Pitch in a Neural Net Model of Chord Classification , 1989 .

[3]  Peter M. Todd,et al.  A Connectionist Approach To Algorithmic Composition , 1989 .

[4]  Michael I. Jordan.  Attractor dynamics and parallelism in a connectionist sequential machine , 1990 .

[5]  Louis P. DiPalma,et al.  Music and Connectionism , 1991 .

[6]  Gerald Tesauro,et al.  TD-Gammon, a Self-Teaching Backgammon Program, Achieves Master-Level Play , 1994, Neural Computation.

[7]  Michael C. Mozer,et al.  Neural Network Music Composition by Prediction: Exploring the Benefits of Psychoacoustic Constraints and Multi-scale Processing , 1994, Connect. Sci..

[8]  David S. Watson,et al.  A Machine Learning Approach to Musical Style Recognition , 1997, ICMC.

[9]  Andrew G. Barto,et al.  Reinforcement learning , 1998 .

[10]  Paolo Campolucci,et al.  A Circuit Theory Approach to Recurrent Neural Network Architectures and Learning Methods , 1998 .

[11]  P. Todd,et al.  Musical networks: Parallel distributed perception and performance , 1999 .

[12]  Jürgen Schmidhuber,et al.  Learning to Forget: Continual Prediction with LSTM , 2000, Neural Computation.

[13]  Bram Bakker,et al.  Reinforcement Learning with Long Short-Term Memory , 2001, NIPS.

[14]  Judy A. Franklin.  Multi-Phase Learning for Jazz Improvisation and Interaction , 2001 .

[15]  Jürgen Schmidhuber,et al.  Learning the Long-Term Structure of the Blues , 2002, ICANN.

[16]  Judy A. Franklin,et al.  Recurrent Neural Networks and Pitch Representations for Music Tasks , 2004, FLAIRS.