A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2023; you can also visit the original URL.
The file type is application/pdf
.
Filters
Exploring Representation Learning for Small-Footprint Keyword Spotting
[article]
2023
pre-print
In this paper, we investigate representation learning for low-resource keyword spotting (KWS). The main challenges of KWS are limited labeled data and limited available device resources. ...
To address those challenges, we explore representation learning for KWS by self-supervised contrastive learning and self-training with pretrained model. ...
In this study, we aim at exploring representation learning to improve the small-footprint KWS model from two aspects. ...
doi:10.21437/interspeech.2022-10558
arXiv:2303.10912v1
fatcat:yhn7qab5dzfyjozkogqqvv7ng4
Attention-based End-to-End Models for Small-Footprint Keyword Spotting
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
Improving Small Footprint Few-shot Keyword Spotting with Supervision on Auxiliary Data
[article]
2023
arXiv
pre-print
a small footprint FS-KWS model. ...
Self-supervised learning has been widely adopted for learning representations from unlabeled data; however, it is known to be suitable for large models with enough capacity and is not practical for training ...
In small-footprint keyword spotting models, learning invariant information with SSL-based pre-training hinders the creation of keyword representations. ...
arXiv:2309.00647v1
fatcat:ovat55ovjjffxaedt3j4wzwnbq
Small Footprint Multi-channel ConvMixer for Keyword Spotting with Centroid Based Awareness
[article]
2022
arXiv
pre-print
It is critical for a keyword spotting model to have a small footprint as it typically runs on-device with low computational resources. ...
In this paper, we present a multi-channel ConvMixer for speech command recognitions. ...
In this paper, we extend the previous work in [21] to build a novel small footprint multi-channel ConvMixer for keyword spotting. ...
arXiv:2204.05445v1
fatcat:2a2eygx3kvbslawgryp3zw2mfe
Convolutional Recurrent Neural Networks for Small-Footprint Keyword Spotting
[article]
2017
arXiv
pre-print
Keyword spotting (KWS) constitutes a major component of human-technology interfaces. ...
Maximizing the detection accuracy at a low false alarm (FA) rate, while minimizing the footprint size, latency and complexity are the goals for KWS. ...
We thank Hui Song for the impulse response measurements used for farfield augmentation. ...
arXiv:1703.05390v3
fatcat:nhprneov4jev5nsgewfmna32wu
Error-diffusion based Speech Feature Quantization for Small-Footprint Keyword Spotting
2022
IEEE Signal Processing Letters
Although small-footprint networks have been widely explored to reduce deployment overhead, low-precision input feature representation still lacks in-depth research. ...
Neural network based keyword spotting (KWS) system is a critical component for user interaction in current smart devices. ...
Meanwhile, a 2-keyword task ("happy" and "stop") is used to simulate a small-footprint KWS application where 3-bit and binary representation are examined. ...
doi:10.1109/lsp.2022.3179208
fatcat:4qqlfs6jknhgldjwlde5ajtrrm
Small-footprint Keyword Spotting with Graph Convolutional Network
[article]
2019
arXiv
pre-print
In this study, we propose a novel context-aware and compact architecture for keyword spotting task. ...
Despite the recent successes of deep neural networks, it remains challenging to achieve high precision keyword spotting task (KWS) on resource-constrained devices. ...
Recently, since the CNN-based methods are widely used in the KWS task, we explore the contextual feature augmentation to help the representations learning for KWS in this work. ...
arXiv:1912.05124v1
fatcat:u2wokgib7bhehneqejhvbf534e
Small-footprint Keyword Spotting Using Deep Neural Network and Connectionist Temporal Classifier
[article]
2017
arXiv
pre-print
Mainly for the sake of solving the lack of keyword-specific data, we propose one Keyword Spotting (KWS) system using Deep Neural Network (DNN) and Connectionist Temporal Classifier (CTC) on power-constrained ...
small-footprint mobile devices, taking full advantage of general corpus from continuous speech recognition which is of great amount. ...
small memory footprint and low computing cost. ...
arXiv:1709.03665v1
fatcat:fmtevd64brcwrfbhpz66g7zpnm
Domain Aware Training for Far-field Small-footprint Keyword Spotting
[article]
2020
arXiv
pre-print
In this paper, we focus on the task of small-footprint keyword spotting under the far-field scenario. ...
To cope with the distortions, we develop three domain aware training systems, including the domain embedding system, the deep CORAL system, and the multi-task learning system. ...
Conclusions In this paper, we concentrate on the task of small-footprint keyword spotting under the far-field environment. ...
arXiv:2005.03633v3
fatcat:af7cezrqlreg5ltl2ciwcaqjju
EdgeSpeechNets: Highly Efficient Deep Neural Networks for Speech Recognition on the Edge
[article]
2018
arXiv
pre-print
Recently, there have been greater efforts in the design of small, low-footprint deep neural networks (DNNs) that are more appropriate for edge devices, with much of the focus on design principles for hand-crafting ...
In this study, we explore a human-machine collaborative design strategy for building low-footprint DNN architectures for speech recognition through a marriage of human-driven principled network design ...
In particular, limited-vocabulary speech recognition [1] , also known as keyword spotting, has recently seen significant interest as an important application of deep learning for mobile, IoT, and other ...
arXiv:1810.08559v2
fatcat:jtsgjllhnjcgpevp7kietnngcq
Efficient Keyword Spotting by capturing long-range interactions with Temporal Lambda Networks
[article]
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. ...
Latter advances reported the residual temporal convolution networks as an effective candidate for small-footprint KWS and have been proposed in [9, 10, 11] to address this task. ...
arXiv:2104.08086v2
fatcat:caw4hjq4rvbnbpnulw4r743jom
Self-supervised speech representation learning for keyword-spotting with light-weight transformers
[article]
2023
arXiv
pre-print
Self-supervised speech representation learning (S3RL) is revolutionizing the way we leverage the ever-growing availability of data. ...
We demonstrate the effectiveness of S3RL on a keyword-spotting (KS) problem by using transformers with 330k parameters and propose a mechanism to enhance utterance-wise distinction, which proves crucial ...
CONCLUSION AND FUTURE WORK In this paper, we explored the feasibility of using self-supervised representation learning on small-footprint models. ...
arXiv:2303.04255v1
fatcat:ya66ycsxknc33bn4houtukz67y
Sequence-to-sequence Models for Small-Footprint Keyword Spotting
[article]
2018
arXiv
pre-print
In this paper, we propose a sequence-to-sequence model for keyword spotting (KWS). ...
Compared with other end-to-end architectures for KWS, our model simplifies the pipelines of production-quality KWS system and satisfies the requirement of high accuracy, low-latency, and small-footprint ...
As a wake-up trigger, KWS should satisfy the requirement of small memory and low CPU footprint, with high accuracy. ...
arXiv:1811.00348v1
fatcat:apqlsmjpgrcbnpdendcttdozey
Learning Efficient Representations for Keyword Spotting with Triplet Loss
[article]
2021
arXiv
pre-print
We fill this gap showing that a combination of two representation learning techniques: a triplet loss-based embedding and a variant of kNN for classification instead of cross-entropy loss significantly ...
(by 26% to 38%) improves the classification accuracy for convolutional networks on a LibriSpeech-derived LibriWords datasets. ...
INTRODUCTION The goal of keyword spotting is to detect a relatively small set of predefined keywords in a stream of user utterances, usually in the context of small-footprint device [1] . ...
arXiv:2101.04792v4
fatcat:b4anqfjirfddrngpbakb56q4va
Deep Spoken Keyword Spotting: An Overview
2021
IEEE Access
INDEX TERMS Keyword spotting, deep learning, acoustic model, small footprint, robustness.
I. ...
Wei, “Small-footprint
of long short-term memory networks for small-footprint keyword spot- keyword spotting with graph convolutional network,” in Proceedings ...
doi:10.1109/access.2021.3139508
fatcat:i4pfpfxcpretlkbefp7owtxcti
« Previous
Showing results 1 — 15 out of 2,240 results