Deep Learning and Music Adversaries

Kereliuk, Corey; Sturm, Bob L.; Larsen, Jan

doi:10.1109/tmm.2015.2478068

Cited by 114 publications

(86 citation statements)

References 43 publications

(107 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Also, we can see that the multi-level and multi-scale aggregation technique generally improves the performance, particularly in GTZAN. [24] 0.632 ---Temporal features [34] 0.659 ---CNN using artist-labels [35] 0 …”

Section: Transfer Learning and Comparison To State-of-the-artsmentioning

confidence: 99%

SampleCNN: End-to-End Deep Convolutional Neural Networks Using Very Small Filters for Music Classification

et al. 2018

View full text Add to dashboard Cite

Convolutional Neural Networks (CNN) have been applied to diverse machine learning tasks for different modalities of raw data in an end-to-end fashion. In the audio domain, a raw waveform-based approach has been explored to directly learn hierarchical characteristics of audio. However, the majority of previous studies have limited their model capacity by taking a frame-level structure similar to short-time Fourier transforms. We previously proposed a CNN architecture which learns representations using sample-level filters beyond typical frame-level input representations. The architecture showed comparable performance to the spectrogram-based CNN model in music auto-tagging. In this paper, we extend the previous work in three ways. First, considering the sample-level model requires much longer training time, we progressively downsample the input signals and examine how it affects the performance. Second, we extend the model using multi-level and multi-scale feature aggregation technique and subsequently conduct transfer learning for several music classification tasks. Finally, we visualize filters learned by the sample-level CNN in each layer to identify hierarchically learned features and show that they are sensitive to log-scaled frequency.

show abstract

Section: Transfer Learning and Comparison To State-of-the-artsmentioning

confidence: 99%

SampleCNN: End-to-End Deep Convolutional Neural Networks Using Very Small Filters for Music Classification

et al. 2018

View full text Add to dashboard Cite

show abstract

“…We used AUC (Area Under Receiver Operating Characteristic) as a primary evaluation metric for music autotagging. In addition, we conducted genre classification tasks, GTZAN [16] (10 genres, fault-filtered split that is designed to avoid the repetition of artist across training/validation/test list [17]) and Tagtraum genre annotations on MSD (15 genres, stratified split with 80% training data of CD2C version) [18], in a transfer learning setting where the pre-trained CNNs with MSD are used as feature extractors.…”

Section: A Datasetsmentioning

confidence: 99%

Multi-Level and Multi-Scale Feature Aggregation Using Pretrained Convolutional Neural Networks for Music Auto-Tagging

Lee

Nam

2017

IEEE Signal Process. Lett.

127

View full text Add to dashboard Cite

Abstract-Music auto-tagging is often handled in a similar manner to image classification by regarding the 2D audio spectrogram as image data. However, music auto-tagging is distinguished from image classification in that the tags are highly diverse and have different levels of abstractions. Considering this issue, we propose a convolutional neural networks (CNN)-based architecture that embraces multi-level and multi-scaled features. The architecture is trained in three steps. First, we conduct supervised feature learning to capture local audio features using a set of CNNs with different input sizes. Second, we extract audio features from each layer of the pre-trained convolutional networks separately and aggregate them altogether given a long audio clip. Finally, we put them into fully-connected networks and make final predictions of the tags. Our experiments show that using the combination of multi-level and multi-scale features is highly effective in music auto-tagging and the proposed method outperforms previous state-of-the-arts on the MagnaTagATune dataset and the Million Song Dataset. We further show that the proposed architecture is useful in transfer learning.

show abstract

“…Thus, it is more convenient for resource management. When adding or deleting resource, it only needs to modify the corresponding memory address, so as to inquire and use resources more intuitively and realize the ore convenient operation [10]. Realization block diagram of test questions management function is as shown in figure 6.…”

Section: Realization Of System Question Bank Management Programmentioning

confidence: 99%

Music Solfeggio Learning Platform Construction and Application

Qiao

Yan

2017

Int. J. Emerg. Technol. Learn.

View full text Add to dashboard Cite

Abstract-To better develop the music learning, this paper completes the design and realization of a music solfeggio teaching system by combining with practical teaching conditions of the music academy. Firstly, it elaborates the main functions needing to be possessed by solfeggio teaching system by starting from actual demands of the users, puts forward overall design scheme of the system, and gives detailed design to main function module and database of the system. Secondly, it analyzes and researches theoretical basis of the solfeggio teaching system design, and proposes the construction scheme of teaching knowledge point repository and question bank system, including solfeggio repository information setting and system paper constructing strategy. It is indicated by the system analysis results that: this platform design provides an effective learning and inspection means to the implementation of solfeggio teaching. Thus, it draws the conclusions that: learning system of this paper can directly serve for course learning of the students majoring in music, and it has important practical significance and application value in promoting development of the music education informationization.

show abstract

Deep Learning and Music Adversaries

Cited by 114 publications

References 43 publications

SampleCNN: End-to-End Deep Convolutional Neural Networks Using Very Small Filters for Music Classification

SampleCNN: End-to-End Deep Convolutional Neural Networks Using Very Small Filters for Music Classification

Multi-Level and Multi-Scale Feature Aggregation Using Pretrained Convolutional Neural Networks for Music Auto-Tagging

Music Solfeggio Learning Platform Construction and Application

Contact Info

Product

Resources

About