A First Look at Music Composition using LSTM Recurrent Neural Networks.
Published 15 March 2002Open access
Douglas Eck, Juergen Schmidhuber
Citations204
Generate an AI Snapshot to get a quick, structured summary of this paper.
Study Snapshot
ObjectiveStudy objective
MethodsResearch methodology
PopulationPopulation studied
Sample sizeSample sizes
OutcomesStudy outcomes here
ResultsStudy results comes here
LimitationsResearch study limitations comes here
A concise AI-generated summary of the paper will appear here once you click Generate AI Snapshot.
TL;DR
Long Short-Term Memory is shown to be able to play the blues with good timing and proper structure as long as one is willing to listen, and once the network has found the relevant structure it does not drift from it.
Abstract
In general music composed by recurrent neural networks (RNNs) su#ers from a lack of global structure. Though networks can learn note-by-note transition probabilities and even reproduce phrases, attempts at learning an entire musical form and using that knowledge to guide composition have been unsuccessful. The reason for this failure seems to be that RNNs cannot keep track of temporally distant events that indicate global music structure.
Keywords
Computer ScienceNeuroscience
Neural ComputationLong Short-Term Memory
98,079 Citations1997Sepp Hochreiter, Jürgen Schmidhuber
A novel, efficient, gradient based method called long short-term memory (LSTM) is introduced, which can learn to bridge minimal time lags in excess of 1000 discrete-time steps by enforcing constant error flow through constant error carousels within special units.
IEEE Transactions on Neural NetworksLearning long-term dependencies with gradient descent is difficult
8,461 Citations1994Yoshua Bengio, P. Simard +1 more
This work shows why gradient based learning algorithms face an increasingly difficult problem as the duration of the dependencies to be captured increases, and exposes a trade-off between efficient learning by gradient descent and latching on information for long periods.
Neural ComputationLearning to Forget: Continual Prediction with LSTM
5,419 Citations2000Felix A. Gers, Jürgen Schmidhuber +1 more
This work identifies a weakness of LSTM networks processing continual input streams that are not a priori segmented into subsequences with explicitly marked ends at which the network's internal state could be reset, and proposes a novel, adaptive forget gate that enables an LSTm cell to learn to reset itself at appropriate times, thus releasing internal resources.
IEEE Transactions on Neural NetworksLSTM recurrent networks learn simple context-free and context-sensitive languages
726 Citations2001Felix A. Gers, E. Schmidhuber
Long short-term memory (LSTM) variants are also the first RNNs to learn a simple context-sensitive language, namely a(n)b( n)c(n).
Recurrent nets that time and count
641 Citations2000Felix A. Gers, Jürgen Schmidhuber
Surprisingly, LSTM augmented by "peephole connections" from its internal cells to its multiplicative gates can learn the fine distinction between sequences of spikes separated by either 50 or 49 discrete time steps, without the help of any short training exemplars.
Gradient Flow in Recurrent Nets: the Difficulty of Learning Long-Term Dependencies
530 Citations2001Sepp Hochreiter, Yoshua Bengio +2 more
Gradient-based learning algorithms for recurrent networks and their computational complexity
416 Citations1995Ronald J. Williams, David Zipser
Computer Music JournalA Connectionist Approach to Algorithmic Composition
215 Citations1989Peter M. Todd
This paper presents a particular type of PDP network for music composition applications and provides an indication of the power and range of P DP methods for algorithmic composition and to encourage others to begin exploring this new approach.
Neural NetworksKalman filters improve LSTM network performance in problems unsolvable by traditional recurrent nets
109 Citations2003Juan Antonio Pérez-Ortiz, Felix A. Gers +2 more
A set of experiments which are unsolvable by classical recurrent networks but which are solved elegantly and robustly and quickly by LSTM combined with Kalman filters are presented.
Computer Music JournalModeling the Perception of Tonal Structure with Neural Nets
98 Citations1989Jamshed J. Bharucha, Peter M. Todd
The notion that individual neurons specialize to respond to complex auditory patterns has some preliminary empirical support from single-cell recording studies on animals (Weinberger and McKenna 1988).
Lecture notes in computer scienceLearning the Long-Term Structure of the Blues
67 Citations2002Douglas Eck, Jürgen Schmidhuber
LSTM is able to play the blues with good timing and proper structure as long as one is willing to listen and once the network has found the relevant structure it does not drift from it.
Psychological ResearchFinding downbeats with a relaxation oscillator
45 Citations2002Douglas Eck
A relaxation oscillator model of neural spiking dynamics is applied to the task of finding downbeats in rhythmical patterns and is compared to other oscillator models, identifying some shortcomings and relating model behavior to dynamical properties of relaxation oscillators.
Computer Music JournalThe Representation of Pitch in a Neural Net Model of Chord Classification
41 Citations1989Bernice Laden, Douglas H. Keefe
Researchers who want to use neural nets for musical signal processing may be interested in both low and high- level representations in order to enable analysis of recorded musical sound and the subsequent conversion into the high-level abstraction of a musical.
International Journal of Uncertainty Fuzziness and Knowledge-Based SystemsSpeeding Up Backpropagation Algorithms by Using Cross-Entropy Combined with Pattern Normalization
38 Citations1998Merten Joost, Wolfram Schiffmann
This paper demonstrates how the backpropagation algorithm (BP) and its variants can be accelerated significantly while the quality of the trained nets will increase.
Lecture notes in computer scienceA Network of Relaxation Oscillators that Finds Downbeats in Rhythms
13 Citations2001Douglas Eck
A network of relaxation oscillators is used to find downbeats in rhythmical patterns and its performance is analyzed to relate its success to previous work dealing with fast synchrony in coupled oscillators.