Exploring convolutional neural network structures and optimization techniques for speech recognition
By Ossama Abdel-Hamid, Li Deng, Dong Yu. 2013
417 citations recorded by the source corpus.
Subfields: Artificial Intelligence, 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 Artificial Intelligence
In Signal Processing
Citations in the corpus
- A Hybrid Differential Evolution Approach to Designing Deep Convolutional Neural Networks for Image Classification (2018, background): “Convolutional Neural Networks (CNNs) have proved their dominating spot in various machine learning tasks, such as speech recognition [1], sentence classifi-cation [6] and image classification [7].”
- Very deep convolutional networks for end-to-end speech recognition (2016, methodology): “Convolutional Neural Networks (CNNs) [8] have been successfully applied to many ASR tasks [9, 10, 11].”
- Large-Scale Deep Learning on the YFCC100M Dataset (2015, background): “Neural networks have been shown to perform well in a wide variety of tasks, including text analysis [4], speech recognition [5, 6, 7], various classification tasks [8, 9], and most notably unsupervised and supervised feature learning on natural imagery [1, 2, 3].”
- Understanding of a convolutional neural network (2017, background): “Thus, if we want to have 32×32 neurons in the next layer, then we will have 5×5×3 by 32x32 connections which is 76,800 connections (compared to 3,145,728 for full connectivity), as in [1,9,13,14,17].”
- Evolving Deep Convolutional Neural Networks by Variable-Length Particle Swarm Optimization for Image Classification (2018, methodology): “With regard to Conv layers, firstly, there are three key parameters - filter size, number of feature maps and stride size listed in the column of Parameter in Table I, which are the fundamental factors affecting the performance of CNNs; Secondly, based on the size of benchmark datasets, the range of”
- Segmental Spatiotemporal CNNs for Fine-Grained Action Segmentation (2016, background): “It is more similar to those in speech recognition (e.g. [35,36]) which learn phonemes using 1D convolutional filters or in robotics which learn sensor-based action filters [37].”
- Convolutional Neural Networks for Distant Speech Recognition (2014, background): “, (1) where (·) is a sigmoid nonlinearity, ⇤ denotes linear valid convolution1, wsj 2 RF⇥Z is a weight vector of the j-th filter acting on the local input usk of the s-th input channel, and bj,k is an additive bias for the j-th filter and k-th convolutional band.”
- Convolutional deep maxout networks for phone recognition (2014, background): “Excellent speech recognition results have lately been reported using deep Convolutional Neural Networks (CNNs) [1, 2, 3, 4, 5, 6].”
- Deep Convolutional Neural Networks with Layer-Wise Context Expansion and Attention (2016, methodology): “Originally, CNNs were applied along the frequency dimension [6, 7, 8, 10] to reduce variations caused by vocal tract length differences between speakers.”
- Angel-Eye: A Complete Design Flow for Mapping CNN Onto Embedded FPGA (2018, methodology): “Some audio algorithms also involves CNN as one of the feature extraction steps [5].”
- Deep Learning in Spiking Neural Networks (2018, background): “n and pooling (sub-sampling) layers followed by a MLP classifier. This type of network has shown outstanding performance in image recognition [2], [135], [136], [137], speech recognition [108], [138], [139], bioinformatics [140], [141], [111], object detection and segmentation [142], [3], and so on. ”
- An introduction to computational networks and the computational network toolkit (invited talk) (2014, background): “International Conference on Neural Networks (IJCNN), 2012, pp. 1–6.”