Recurrent Neural Networks for Music Computation

Some researchers in the computational sciences have considered music computation, including music reproduction and generation, as a dynamic system, i.e., a feedback process. The key element is that the state of the musical system depends on a history of past states. Recurrent (neural) networks have been deployed as models for learning musical processes. We first present a tutorial discussion of recurrent networks, covering those that have been used for music learning. Following this, we examine a thread of development of these recurrent networks for music computation that shows how more intricate music has been learned as the state of the art in recurrent networks improves. We present our findings that show that a long short-term memory recurrent network, with new representations that include music knowledge, can learn musical tasks, and can learn to reproduce long songs. Then, given a reharmonization of the chordal structure, it can generate an improvisation.

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

[2]  G. Miller.  Learning to Forget , 2004, Science.

[3]  Jing Peng,et al.  An Efficient Gradient-Based Algorithm for On-Line Training of Recurrent Network Trajectories , 1990, Neural Computation.

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

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

[6]  Judy A. Franklin,et al.  Predicting Reinforcement of Pitch Sequences via LSTM and TD , 2004, ICMC.

[7]  Peter Desain,et al.  A Connectionist Quantizer , 1989, ICMC.

[8]  Danilo P. Mandic,et al.  Recurrent Neural Networks for Prediction , 2001 .

[9]  Jürgen Schmidhuber,et al.  Learning to forget: continual prediction with LSTM , 1999 .

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

[11]  Roger B. Dannenberg,et al.  Tracking Musical Beats in Real Time , 1990, ICMC.

[12]  R. Shepard,et al.  Toward a universal law of generalization for psychological science. , 1987, Science.

[13]  Mark Steedman,et al.  The well-tempered computer , 1994, Philosophical Transactions of the Royal Society of London. Series A: Physical and Engineering Sciences.

[14]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

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

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

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

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

[19]  Yoshua Bengio,et al.  Gradient Flow in Recurrent Nets: the Difficulty of Learning Long-Term Dependencies , 2001 .

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

[21]  John F. Kolen,et al.  Dynamical Recurrent Networks , 2001 .

[22]  John F. Kolen,et al.  Resonance and the Perception of Musical Meter , 1994, Connect. Sci..

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

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

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