A deep convolutional neural network using heterogeneous pooling for trading acoustic invariance with phonetic confusion

By Li Deng, Ossama Abdel-Hamid, Dong Yu. 2013

184 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

Deep Learning: Methods and Applications (2014, methodology): “It was not until recently that researchers have discovered that the timedimension invariance is less important than the frequency-dimension invariance for speech recognition [1, 3, 81].”
Modeling Interestingness with Deep Neural Networks (2014, positive): “These keywords represent the entities or concepts that might interest users, via the convolutional and max-pooling layers which are related to the deep models used for computer vision (Krizhevsky et al., 2013) and speech recognition (Deng et al., 2013a) but are not used in Huang et al.’s model.”
A tutorial survey of architectures, algorithms, and applications for deep learning (2014, methodology): “Recently, the deep learning technique was successfully applied to phone recognition [23, 24, 116, 126, 165, 166] and large vocabulary speech recognition tasks [22, 25, 135, 155, 156, 167–169] by integrating the powerful discriminative training ability of DNNs with the sequential modeling ability of ”
A Primal-Dual Method for Training Recurrent Neural Networks Constrained by the Echo-State Property (2013, methodology): “…deep neural networks (DNN) that drastically cut errors in large vocabulary speech recognition (Yu et al., 2010; Dahl et al., 2011; Seide et al., 2011; Dahl et al., 2012; Kingsbury et al., 2012; Hinton et al., 2012; Deng et al., 2013b,c; Yu et al., 2013; Dahl et al., 2013; Sainath et al., 2013).”
Recent advances in deep learning for speech research at Microsoft (2013, unknown): “Details of this new deep CNN and error analysis are provided in [17].”
Dropout Improves Recurrent Neural Networks for Handwriting Recognition (2013, methodology): “It is shown to be effective with deep convolutional networks [12], [13], [14], feed-forward networks [15], [16], [17] but, to the best of our knowledge, has never been applied to RNNs.”
Environmental sound classification with convolutional neural networks (2015, methodology): “Although primarily used in visual recognition contexts, convolutional architectures have been also successfully applied in speech [12–18] and music analysis [19, 20].”
Exploring convolutional neural network structures and optimization techniques for speech recognition (2013, background): “Moreover, in [13], a heterogeneous pooling structure proved to be beneﬁcial.”
MoDeep: A Deep Learning Framework Using Motion Features for Human Pose Estimation (2014, background): “Deep Learning based Techniques: Recently, state-of-the-art performance has been reported on many vision tasks using deep learning algorithms [43, 44, 45, 46, 47, 48].”
Learning filter banks within a deep neural network framework (2013, background): “[18] introduces a varied pooling strategy for each weight that spans across a localized frequency region (i.”
Convolutional Neural Networks for Speech Recognition (2014, background): “There have been extensions of the work described in this paper to other larger vocabulary speech recognition tasks that lend further support to the value of this approach [39], [40].”
Deep Learning over Multi-field Categorical Data - - A Case Study on User Response Prediction (2016, methodology): “DNNs have been successfully applied in computer vision [40], speech recognition [8] and natural language processing (NLP) [7,19,33].”
Phone recognition with hierarchical convolutional deep maxout networks (2015, background): “improved their results by combining various pooling sizes within the same system [8].”
Deep segmental neural networks for speech recognition (2013, background): “Recently, deep-neural-network hidden Markov model (DNN/HMM) hybrid systems have achieved remarkable performance in many large vocabulary speech recognition tasks [1, 2, 3, 4, 5, 6, 7, 8].”
Maximum a posteriori Based Decoding for CTC Acoustic Models (2016, background): “Encouraged by the strong performance of DNNs, various extensions of the network architecture have been proposed, such as convolutional neural networks (CNNs) [6], sigmoid-unit-type recurrent neural networks (RNNs) [7, 8, 9, 10], and long short-term memory neural networks [11, 12, 13, 14, 15].”