Filters








225 Hits in 3.0 sec

Convolutional Recurrent Neural Networks for Small-Footprint Keyword Spotting [article]

Sercan O. Arik, Markus Kliegl, Rewon Child, Joel Hestness, Andrew Gibiansky, Chris Fougner, Ryan Prenger, Adam Coates
2017 arXiv   pre-print
Towards achieving them, we study Convolutional Recurrent Neural Networks (CRNNs).  ...  Keyword spotting (KWS) constitutes a major component of human-technology interfaces.  ...  We thank Hui Song for the impulse response measurements used for farfield augmentation.  ... 
arXiv:1703.05390v3 fatcat:nhprneov4jev5nsgewfmna32wu

Convolutional Recurrent Neural Networks for Small-Footprint Keyword Spotting

Sercan Ö. Arık, Markus Kliegl, Rewon Child, Joel Hestness, Andrew Gibiansky, Chris Fougner, Ryan Prenger, Adam Coates
2017 Interspeech 2017   unpublished
Towards achieving them, we study Convolutional Recurrent Neural Networks (CRNNs).  ...  Keyword spotting (KWS) constitutes a major component of human-technology interfaces.  ...  We thank Hui Song for the impulse response measurements used for far-field augmentation.  ... 
doi:10.21437/interspeech.2017-1737 fatcat:qvlvhd3karhtzpcc3v3bjolcnu

Deep Residual Learning for Small-Footprint Keyword Spotting [article]

Raphael Tang, Jimmy Lin
2018 arXiv   pre-print
Our best residual network (ResNet) implementation significantly outperforms Google's previous convolutional neural networks in terms of accuracy.  ...  By varying model depth and width, we can achieve compact models that also outperform previous small-footprint variants.  ...  In recent years, neural networks have been shown to provide effective solutions to the small-footprint keyword spotting problem.  ... 
arXiv:1710.10361v2 fatcat:kkhpormgljf27a63wqw5okchna

Hello Edge: Keyword Spotting on Microcontrollers [article]

Yundong Zhang, Naveen Suda, Liangzhen Lai, Vikas Chandra
2018 arXiv   pre-print
We train various neural network architectures for keyword spotting published in literature to compare their accuracy and memory/compute requirements.  ...  We further explore the depthwise separable convolutional neural network (DS-CNN) and compare it against other neural network architectures.  ...  We would also like to thank Pete Warden from Google's TensorFlow team for his valuable inputs and feedback on this project.  ... 
arXiv:1711.07128v3 fatcat:swrltzaqc5hvjay7ofrx3r4lwy

An Experimental Analysis of the Power Consumption of Convolutional Neural Networks for Keyword Spotting [article]

Raphael Tang, Weijie Wang, Zhucheng Tu, Jimmy Lin
2018 arXiv   pre-print
Nearly all previous work on small-footprint keyword spotting with neural networks quantify model footprint in terms of the number of parameters and multiply operations for a feedforward inference pass.  ...  In this paper, we study the power consumption of a family of convolutional neural networks for keyword spotting on a Raspberry Pi.  ...  Despite more recent work in applying recurrent neural networks to the keyword spotting task [3, 4] , we focus on the family of CNN models for several reasons.  ... 
arXiv:1711.00333v2 fatcat:zpnmfqdb7jglxe4vt6tzyeanbu

Attention-based End-to-End Models for Small-Footprint Keyword Spotting

Changhao Shan, Junbo Zhang, Yujun Wang, Lei Xie
2018 Interspeech 2018  
In this paper, we propose an attention-based end-to-end neural approach for small-footprint keyword spotting (KWS), which aims to simplify the pipelines of building a production-quality KWS system.  ...  Finally, by linear transformation and softmax function, the vector becomes a score used for keyword detection.  ...  Acknowledgements The authors would like to thank Jingyong Hou for helpful comments and suggestions.  ... 
doi:10.21437/interspeech.2018-1777 dblp:conf/interspeech/ShanZWX18 fatcat:tefhrrsnvndwvirmh2dug6waxy

Predicting detection filters for small footprint open-vocabulary keyword spotting [article]

Theodore Bluche, Thibault Gisselbrecht
2020 arXiv   pre-print
a detector for any keyword.  ...  We also propose a method to fine-tune the model when specific training data is available for some keywords, which yields a performance similar to a standard speech command neural network while keeping  ...  cs.CL] 29 Sep 2020 Keyword spotting neural network The neural network is made of a stack of unidirectional LSTM layers, followed by two convolutional layers.  ... 
arXiv:1912.07575v2 fatcat:2zjnljraengx5phaegwjzirlee

Small-footprint Keyword Spotting with Graph Convolutional Network [article]

Xi Chen, Shouyi Yin, Dandan Song, Peng Ouyang, Leibo Liu, Shaojun Wei
2019 arXiv   pre-print
Despite the recent successes of deep neural networks, it remains challenging to achieve high precision keyword spotting task (KWS) on resource-constrained devices.  ...  In this study, we propose a novel context-aware and compact architecture for keyword spotting task.  ...  METHOD We describe our method for building a compact and efficient network to achieve the small footprint KWS system.  ... 
arXiv:1912.05124v1 fatcat:u2wokgib7bhehneqejhvbf534e

Predicting Detection Filters for Small Footprint Open-Vocabulary Keyword Spotting

Théodore Bluche, Thibault Gisselbrecht
2020 Interspeech 2020  
Index Terms: speech recognition, keyword spotting, neural networks 2.  ...  Keyword spotting neural network When keywords to detect are known in advance, and when training data containing those keywords are available, a neural network can be trained in an end-to-end fashion to  ...  Keyword spotting neural network The neural network is made of a stack of unidirectional LSTM layers, followed by two convolutional layers.  ... 
doi:10.21437/interspeech.2020-1186 dblp:conf/interspeech/BlucheG20 fatcat:ra3blq66abdnbiflhdom3foayy

Efficient keyword spotting using dilated convolutions and gating [article]

Alice Coucke, Mohammed Chlieh, Thibault Gisselbrecht, David Leroy, Mathieu Poumeyrol, Thibaut Lavril
2019 arXiv   pre-print
We explore the application of end-to-end stateless temporal modeling to small-footprint keyword spotting as opposed to recurrent networks that model long-term temporal dependencies using internal states  ...  Our experimental results show that our model outperforms a max-pooling loss trained recurrent neural network using LSTM cells, with a significant decrease in false rejection rate.  ...  We are indebted to the users of the Snips Voice Platform for valuable feedback.  ... 
arXiv:1811.07684v2 fatcat:xyb3cvhn45gbjcoyqfvdxib3te

Domain Aware Training for Far-field Small-footprint Keyword Spotting [article]

Haiwei Wu, Yan Jia, Yuanfei Nie, Ming Li
2020 arXiv   pre-print
In this paper, we focus on the task of small-footprint keyword spotting under the far-field scenario.  ...  Our baseline system is built on the convolutional neural network trained with pooled data of both far-field and close-talking speech.  ...  As for modeling, many structures based on Convolutional Neural Network (CNN) [17] , Recurrent Neural Network (RNN), Convolutional Recurrent Neural Network (CRNN) [1] , Long Short Time Memory [21] (  ... 
arXiv:2005.03633v3 fatcat:af7cezrqlreg5ltl2ciwcaqjju

Encoder-Decoder Neural Architecture Optimization for Keyword Spotting [article]

Tong Mo, Bang Liu
2021 arXiv   pre-print
Keyword spotting aims to identify specific keyword audio utterances. In recent years, deep convolutional neural networks have been widely utilized in keyword spotting systems.  ...  In this paper, we utilize neural architecture search to design convolutional neural network models that can boost the performance of keyword spotting while maintaining an acceptable memory footprint.  ...  [1] introduced CNNs into KWS and showed that CNNs performed well on small footprint keyword spotting.  ... 
arXiv:2106.02738v1 fatcat:ekl3ju5g4bfxdcry62kam3v6kq

Efficient Keyword Spotting by capturing long-range interactions with Temporal Lambda Networks [article]

Biel Tura, Santiago Escuder, Ferran Diego, Carlos Segura, Jordi Luque
2021 arXiv   pre-print
However, they are computationally expensive and unnecessarily complex for keyword spotting, a task targeted to small-footprint devices.  ...  This work explores the application of Lambda networks, an alternative framework for capturing long-range interactions without attention, for the keyword spotting task.  ...  The usage of recurrent neural networks for the keyword spotting task has also been studied in the work by Arik et al. [20] .  ... 
arXiv:2104.08086v2 fatcat:caw4hjq4rvbnbpnulw4r743jom

Tiny-CRNN: Streaming Wakeword Detection In A Low Footprint Setting [article]

Mohammad Omar Khursheed, Christin Jose, Rajath Kumar, Gengshen Fu, Brian Kulis, Santosh Kumar Cheekatmalla
2021 arXiv   pre-print
In this work, we propose Tiny-CRNN (Tiny Convolutional Recurrent Neural Network) models applied to the problem of wakeword detection, and augment them with scaled dot product attention.  ...  We find that, compared to Convolutional Neural Network models, False Accepts in a 250k parameter budget can be reduced by 25% with a 10% reduction in parameter size by using models based on the Tiny-CRNN  ...  Recurrent neural networks (RNNs), a class of neural networks used to process sequential data, have also been found to be useful for keyword spotting tasks [7] , which is furthered through the use of recurrent  ... 
arXiv:2109.14725v1 fatcat:i7brm3rt7va3tk6zz2og5bu2di

Lightweight dynamic filter for keyword spotting [article]

Donghyeon Kim, Kyungdeuk Ko, Jeong-gi Kwak, David K. Han, Hanseok Ko
2021 arXiv   pre-print
Keyword Spotting (KWS) from speech signal is widely applied for being fully hands free speech recognition. The KWS network is designed as a small footprint model to be constantly monitored.  ...  The experiments show that our model is robustly working on unseen noise and small training data environment by using small computational resource.  ...  Additionally, Convolutional Recurrent Neural Networks (CRNN) based models have been applied to combine the strengths of CNN and Recurrent Neural Network (RNN) [5] .  ... 
arXiv:2109.11165v3 fatcat:ma2igcbkqncm7dgunu3q4gfs24
« Previous Showing results 1 — 15 out of 225 results