Unsupervised Filterbank Learning Using Convolutional Restricted Boltzmann Machine for Environmental Sound Classification

By Hardik B. Sailor, Dharmesh M. Agrawal, Hemant A. Patil. 2017

97 citations recorded by the source corpus.

Subfields: Biomedical Engineering, Natural Language Processing, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Biomedical Engineering

1. Andrew Zisserman
2. Yuxuan Wang
3. Jean-Baptiste Alayrac

In Natural Language Processing

1. Andreas R. Dengel
2. R. Schneider
3. Chang-Dong Yoo
4. Lucas Smaira
5. Andrew Zisserman
6. Yuxuan Wang
7. Jean-Baptiste Alayrac
8. Relja Arandjelovi'c

In Signal Processing

1. Andrew Zisserman
2. Jean-Baptiste Alayrac
3. Adrià Recasens
4. Jeffrey De Fauw
5. Lucas Smaira
6. Relja Arandjelovi'c
7. Qiuqiang Kong
8. Sander Dieleman

Citations in the corpus

• Self-Supervised MultiModal Versatile Networks (2020, methodology): “Under that challenging comparison, MMV performance on HMDB51 and UCF101 is getting close to the best supervised method that leverage both ImageNet and Kinetics [90], while on ESC-50 it is even better than the best supervised result [74] by 1.”
• Cooperative Learning of Audio and Video Models from Self-Supervised Synchronization (2018, neutral): “Human performance [21] n/a n/a n/a 81.3 - State-of-the-art (RBM)[31] none none none 86.5 -”
• Audio-Visual Instance Discrimination with Cross-Modal Agreement (2020, neutral): “ConvRBM [68] 86.”
• PANNs: Large-Scale Pretrained Audio Neural Networks for Audio Pattern Recognition (2019, background): “, [49] proposed a state-of-theart system for ESC-50, achieved an accuracy of 0.”
• Self-Supervised Learning by Cross-Modal Audio-Video Clustering (2019, unknown): “1% gap with [56].”
• CNN-Based Learnable Gammatone Filterbank and Equal-Loudness Normalization for Environmental Sound Classification (2020, background): “Supervised [15], [16], [18]–[21] and unsupervised [17] approaches have been proposed for ESC.”
• AST: Audio Spectrogram Transformer (2021, unknown): “The current best results on ESC-50 are 86.5% accuracy (trained from scratch, SOTA-S) [33] and 94.7% accuracy (with AudioSet pretraining, SOTA-P) [7].”
• Audioclip: Extending Clip to Image, Text and Audio (2021, methodology): “To successfully solve this task, different approaches were proposed that included the use of one- [27,28] or two-dimensional Convolutional Neural Networks (CNN) operating on static [18,24,32,9,15,17,33,8,30] or trainable [23,10] time-frequency transformation of raw audio.”
• Co-Training of Audio and Video Representations from Self-Co-Training of Audio and Video Representations from Self-Supervised Temporal Synchronization Supervised Temporal Synchronization (2018, unknown): “3% State-of-the-art (RBM) [20] none none 86.”