Cover song identification is an important problem in the field of Music Information Retrieval. Most existing methods rely on hand-crafted features and sequence alignment methods, and further breakthrough is hard to achieve. In this paper, Convolutional Neural Networks (CNNs) are used for representation learning toward this task. We show that they could be naturally adapted to deal with key transposition in cover songs. Additionally, Temporal Pyramid Pooling is utilized to extract information on different scales and transform songs with different lengths into fixed-dimensional representations. Furthermore, a training scheme is designed to enhance the robustness of our model. Extensive experiments demonstrate that combined with these techniques, our approach is robust against musical variations existing in cover songs and outperforms state-of-the-art methods on several datasets with low time complexity.

No abstract

BackgroundThe field of psychiatry has seen significant progress in recent years due to worldwide contributions. National productivity, however, in the field of psychiatry is still unclear. In our study, we investigated contributions of individual nations to the field of psychiatry.MethodsThe Web of Science was used to perform a search from 2011 to 2015 on the subject category “psychiatry”. The total number of articles, citations and the per capita numbers were obtained to analyze the contributions of different countries.ResultsIn psychiatry journals from 2011 to 2015, 84,760 articles were published worldwide. The most productive world areas were North America, East Asia, Europe and Oceania. The percentage of articles published in high-income countries was 87.77%, middle-income countries published 12.07%, and lower-income published 0.16%. Most articles were published by the United States (32.68%); the United Kingdom was next (8.59%), which was followed by Germany (6.77%), Australia (5.87%), and Canada (4.9%). The country with the highest number of citations (243,394) was the United States. A positive correlation was found between the population/GDP and the number of publications (P < 0.01). Australia ranked the highest when normalized to population size, and the Netherlands and Norway were next. The Netherlands ranked highest, followed by Israel and Australia when adjusted for GDP.ConclusionsThe authorship of most of the psychiatry articles was from high-income countries and few papers came from low-income countries. The most productive country was the United States. However, when normalized to population size and GDP, some European and Oceania countries were most productive.Electronic supplementary materialThe online version of this article (doi:10.1186/s13033-017-0127-5) contains supplementary material, which is available to authorized users.