资源分类
》 资源 》论文》Neural Dynamic Programming for Musical Self Similarity

Neural Dynamic Programming for Musical Self Similarity

2020-03-16 | |  137 |   114 |   0

Abstract

We present a neural sequence model designed specifically for symbolic music. The model is based on a learned edit distance mechanism which generalises a classic recursion from computer science, leading to a neural dynamic program. Repeated motifs are detected by learning the transfo mations between them. We represent the arising computational dependencies using a novel data structure, the edit tree; this perspective suggest natural approximations which afford the scaling up of our otherwise cubic time algorithm. We demonstrate our model on real and synthetic data; in all cases it out-performs a strong stacked long short-term memory benchmark.

上一篇：On the Limitations of First-Order Approximation in GAN Dynamics

下一篇：K-means clustering using random matrix sparsification

用户评价
全部评价

热门资源

  • Regularizing RNNs...

    Recently, caption generation with an encoder-de...

  • Deep Cross-media ...

    Cross-media retrieval is a research hotspot in ...

  • Learning Expressi...

    Facial expression is temporally dynamic event w...

  • Compact MDDs for ...

    Pseudo-Boolean (PB) constraints are usually en...

  • Attributed Graph ...

    Graph clustering is a fundamental task which di...