Bottom-up broadcast neural network for music genre classification

Liu, Caifeng; Feng, Lin; Liu, Guochao; Wang, Huibing; Liu, Shenglan

doi:10.1007/s11042-020-09643-6

Cited by 85 publications

(69 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The aim of Liu et al's BBNN architecture was to handle the multi-scale of audio feature and use the low level along with high-level information of Mel-spectrogram to achieve higher music genre classification accuracy [1]. The BBNN, equipped with a novel BM module consisting of inception blocks, helped the architecture to handle multi-scale of audio features.…”

Section: Outcome Analysis and Discussionmentioning

confidence: 99%

“…Copyright © 2021 MECS I.J. Information Technology and Computer Science, 2021, 2, 1-14 To us, other than the staggering classification accuracy of BBNN, another significant achievement of the BBNN [1] was the reduction of the need for data-augmentation with the help of its compact parameters, which is often a requirement in traditional CNN. MIR tasks involve training of classifiers from few labeled data are often a challenge, which the BBNN architecture addressed and solved to a certain extent that it eliminates the need for pre-training on bigger datasets such as the Extended Ballroom dataset.…”

Section: Outcome Analysis and Discussionmentioning

confidence: 99%

“…But a major setback for most of the CNN-based music classification models is the requirement of data-augmentation and large datasets for the training of the models, as the models often require to learn large parameters. A study in 2019 solved this problem to a certain extent by the introduction of a novel CNN architecture named Bottom-up Broadcast Neural Network (BBNN) [1]. The architecture was designed to handle multi-scale of audio features and preserve lower-level features, which was usually lost in the previous architectures even though it contained critical information.…”

Section: Literature Reviewmentioning

confidence: 99%

“…In this study, we reviewed and discussed BBNN [1], IndRNN [2], and CRNN-TF [3] three improved architectures of CNN, RNN, and CRNN respectively, that previously used to classify genres of relatively longer clips ( 29 30  seconds) and achieved remarkable results. It is to be noted, as we focused on a particular niche of MIR tasks, that is, on music genre classification, we focused on the use of CRNN-TF for the purpose and ignored the use of the architecture for music auto-tagging, it's transfer-ability from the work of Wang et al [3].…”

Section: Introductionmentioning

confidence: 99%

“…Configuration of the used BBNN (Each convolution layer shown in the table corresponds to the sequence BN-ReLU-Conv), sourced from[1] …”

mentioning

confidence: 99%

See 4 more Smart Citations

Comparative Analysis of Three Improved Deep Learning Architectures for Music Genre Classification

Rafi¹,

Mohammed²,

Prodhan³

et al. 2021

IJITCS

View full text Add to dashboard Cite

Among the many music information retrieval (MIR) tasks, music genre classification is noteworthy. The categorization of music into different groups that came to existence through a complex interplay of cultures, musicians, and various market forces to characterize similarities between compositions and organize collections is known as a music genre. The past researchers extracted various hand-crafted features and developed classifiers based on them. But the major drawback of this approach was the requirement of field expertise. However, in recent times researchers, because of the remarkable classification accuracy of deep learning models, have used similar models for MIR tasks. Convolutional Neural Net- work (CNN), Recurrent Neural Network (RNN), and the hybrid model, Convolutional - Recurrent Neural Network (CRNN), are such prominently used deep learning models for music genre classification along with other MIR tasks and various architectures of these models have achieved state-of-the-art results. In this study, we review and discuss three such architectures of deep learning models, already used for music genre classification of music tracks of length of 29-30 seconds. In particular, we analyze improved CNN, RNN, and CRNN architectures named Bottom-up Broadcast Neural Network (BBNN) [1], Independent Recurrent Neural Network (IndRNN) [2] and CRNN in Time and Frequency dimensions (CRNN- TF) [3] respectively, almost all of the architectures achieved the highest classification accuracy among the variants of their base deep learning model. Hence, this study holds a comparative analysis of the three most impressive architectural variants of the main deep learning models that are prominently used to classify music genre and presents the three architecture, hence the models (CNN, RNN, and CRNN) in one study. We also propose two ways that can improve the performances of the RNN (IndRNN) and CRNN (CRNN-TF) architectures.

show abstract

Section: Outcome Analysis and Discussionmentioning

confidence: 99%

Section: Outcome Analysis and Discussionmentioning

confidence: 99%

Section: Literature Reviewmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

“…Configuration of the used BBNN (Each convolution layer shown in the table corresponds to the sequence BN-ReLU-Conv), sourced from[1] …”

mentioning

confidence: 99%

See 3 more Smart Citations

Comparative Analysis of Three Improved Deep Learning Architectures for Music Genre Classification

Rafi¹,

Mohammed²,

Prodhan³

et al. 2021

IJITCS

View full text Add to dashboard Cite

show abstract

Optimal classification of music genres based on acoustic and visual features

Kumaraswamy¹

2022

Concurrency and Computation

View full text Add to dashboard Cite

Summary Generally, music includes numerous genres like jazz, pop and so forth. Every genre encompasses diverse musical instruments. For classifying music manually, the individual should pay attention to the song and choose the genre. It consumes more time and the individual should know diverse genres. This article aims to introduce an innovative music genre classification model by considering 2 phases. Initially, short‐time Fourier transform, pitch, Timbre, MNMF, and proposed LBP and bag of words‐based features are extracted. Subsequently, the extracted features are subjected to ensemble classifiers, where support vector machine, neural network, random forest are deployed. The output attained from ensemble classifiers is then classified using deep belief network (DBN) which classifies the music genres in a precise way. For precise detection, the training of DBN is made optimal via tuning the weights. This optimization is carried out by a Whale Integrated Sea Lion Algorithm. At last, the supremacy of the developed model is examined via the evaluation of extant techniques.

show abstract

Trajectories of sentiment in 11,816 psychoactive narratives

Friedman,

Ballentine

2023

Human Psychopharmacology

View full text Add to dashboard Cite

ObjectiveCan machine learning (ML) enable data‐driven discovery of how changes in sentiment correlate with different psychoactive experiences? We investigate by training models directly on text testimonials from a diverse 52‐drug pharmacopeia.MethodsUsing large language models (i.e. BERT) and 11,816 publicly‐available testimonials, we predicted 28‐dimensions of sentiment across each narrative, and then validated these predictions with adjudication by a clinical psychiatrist. BERT was then fine‐tuned to predict biochemical and demographic information from these narratives. Lastly, canonical correlation analysis linked the drugs' receptor affinities with word usage, revealing 11 statistically‐significant latent receptor‐experience factors, each mapped to a 3D cortical Atlas.ResultsThese methods elucidate a neurobiologically‐informed, sequence‐sensitive portrait of drug‐induced subjective experiences. The models' results converged, revealing a pervasive distinction between the universal psychedelic heights of feeling in contrast to the grim, mundane, and personal experiences of addiction and mental illness. Notably, MDMA was linked to “Love”, DMT and 5‐MeO‐DMT to “Mystical Experiences” and “Entities and Beings”, and other tryptamines to “Surprise”, “Curiosity” and “Realization".ConclusionsML methods can create unified and robust quantifications of subjective experiences with many different psychoactive substances and timescales. The representations learned are evocative and mutually confirmatory, indicating great potential for ML in characterizing psychoactivity.

show abstract

Bottom-up broadcast neural network for music genre classification

Cited by 85 publications

References 30 publications

Comparative Analysis of Three Improved Deep Learning Architectures for Music Genre Classification

Comparative Analysis of Three Improved Deep Learning Architectures for Music Genre Classification

Optimal classification of music genres based on acoustic and visual features

Trajectories of sentiment in 11,816 psychoactive narratives

Contact Info

Product

Resources

About