paper_id
stringlengths
15
35
title
stringlengths
26
182
authors
listlengths
1
25
isca_url
stringlengths
66
86
pdf_url
stringlengths
65
85
doi
stringlengths
27
30
pages
stringlengths
3
9
bibtex
large_stringlengths
294
850
abstract
large_stringlengths
247
1.59k
arxiv_id
stringlengths
10
10
arxiv_id_source
stringclasses
2 values
swaminathan21_interspeech
CoDERT: Distilling Encoder Representations with Co-Learning for Transducer-Based Speech Recognition
[ "Rupak Vignesh Swaminathan", "Brian King", "Grant P. Strimel", "Jasha Droppo", "Athanasios Mouchtaris" ]
https://www.isca-archive.org/interspeech_2021/swaminathan21_interspeech.html
https://www.isca-archive.org/interspeech_2021/swaminathan21_interspeech.pdf
10.21437/Interspeech.2021-797
4543-4547
@inproceedings{swaminathan21_interspeech, title = {{CoDERT: Distilling Encoder Representations with Co-Learning for Transducer-Based Speech Recognition}}, author = {Rupak Vignesh Swaminathan and Brian King and Grant P. Strimel and Jasha Droppo and Athanasios Mouchtaris}, year = {2021}, booktitle = {...
We propose a simple yet effective method to compress an RNN-Transducer (RNN-T) through the well-known knowledge distillation paradigm. We show that the transducer’s encoder outputs naturally have a high entropy and contain rich information about acoustically similar word-piece confusions. This rich information is suppr...
2106.07734
title_snapshot
gao21f_interspeech
Extremely Low Footprint End-to-End ASR System for Smart Device
[ "Zhifu Gao", "Yiwu Yao", "Shiliang Zhang", "Jun Yang", "Ming Lei", "Ian McLoughlin" ]
https://www.isca-archive.org/interspeech_2021/gao21f_interspeech.html
https://www.isca-archive.org/interspeech_2021/gao21f_interspeech.pdf
10.21437/Interspeech.2021-819
4548-4552
@inproceedings{gao21f_interspeech, title = {{Extremely Low Footprint End-to-End ASR System for Smart Device}}, author = {Zhifu Gao and Yiwu Yao and Shiliang Zhang and Jun Yang and Ming Lei and Ian McLoughlin}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4548--4552}, doi =...
Recently, end-to-end (E2E) speech recognition has become popular, since it can integrate the acoustic, pronunciation and language models into a single neural network, which outperforms conventional models. Among E2E approaches, attention-based models, e.g. Transformer, have emerged as being superior. Such models have o...
2104.05784
title_snapshot
shangguan21_interspeech
Dissecting User-Perceived Latency of On-Device E2E Speech Recognition
[ "Yuan Shangguan", "Rohit Prabhavalkar", "Hang Su", "Jay Mahadeokar", "Yangyang Shi", "Jiatong Zhou", "Chunyang Wu", "Duc Le", "Ozlem Kalinli", "Christian Fuegen", "Michael L. Seltzer" ]
https://www.isca-archive.org/interspeech_2021/shangguan21_interspeech.html
https://www.isca-archive.org/interspeech_2021/shangguan21_interspeech.pdf
10.21437/Interspeech.2021-1887
4553-4557
@inproceedings{shangguan21_interspeech, title = {{Dissecting User-Perceived Latency of On-Device E2E Speech Recognition}}, author = {Yuan Shangguan and Rohit Prabhavalkar and Hang Su and Jay Mahadeokar and Yangyang Shi and Jiatong Zhou and Chunyang Wu and Duc Le and Ozlem Kalinli and Christian Fuegen and Mic...
As speech-enabled devices such as smartphones and smart speakers become increasingly ubiquitous, there is growing interest in building automatic speech recognition (ASR) systems that can run directly on-device; end-to-end (E2E) speech recognition models such as recurrent neural network transducers and their variants ha...
2104.02207
title_snapshot
macoskey21b_interspeech
Amortized Neural Networks for Low-Latency Speech Recognition
[ "Jonathan Macoskey", "Grant P. Strimel", "Jinru Su", "Ariya Rastrow" ]
https://www.isca-archive.org/interspeech_2021/macoskey21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/macoskey21b_interspeech.pdf
10.21437/Interspeech.2021-712
4558-4562
@inproceedings{macoskey21b_interspeech, title = {{Amortized Neural Networks for Low-Latency Speech Recognition}}, author = {Jonathan Macoskey and Grant P. Strimel and Jinru Su and Ariya Rastrow}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4558--4562}, doi = {10.21437/Int...
We introduce Amortized Neural Networks (AmNets), a compute cost- and latency-aware network architecture particularly well-suited for sequence modeling tasks. We apply AmNets to the Recurrent Neural Network Transducer (RNN-T) to reduce compute cost and latency for an automatic speech recognition (ASR) task. The AmNets R...
2108.01553
title_snapshot
botros21_interspeech
Tied & Reduced RNN-T Decoder
[ "Rami Botros", "Tara N. Sainath", "Robert David", "Emmanuel Guzman", "Wei Li", "Yanzhang He" ]
https://www.isca-archive.org/interspeech_2021/botros21_interspeech.html
https://www.isca-archive.org/interspeech_2021/botros21_interspeech.pdf
10.21437/Interspeech.2021-212
4563-4567
@inproceedings{botros21_interspeech, title = {{Tied & Reduced RNN-T Decoder}}, author = {Rami Botros and Tara N. Sainath and Robert David and Emmanuel Guzman and Wei Li and Yanzhang He}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4563--4567}, doi = {10.21437/Interspeech....
Previous works on the Recurrent Neural Network-Transducer (RNN-T) models have shown that, under some conditions, it is possible to simplify its prediction network with little or no loss in recognition accuracy [1, 2, 3]. This is done by limiting the context size of previous labels and/or using a simpler architecture fo...
2109.07513
title_snapshot
kim21m_interspeech
PQK: Model Compression via Pruning, Quantization, and Knowledge Distillation
[ "Jangho Kim", "Simyung Chang", "Nojun Kwak" ]
https://www.isca-archive.org/interspeech_2021/kim21m_interspeech.html
https://www.isca-archive.org/interspeech_2021/kim21m_interspeech.pdf
10.21437/Interspeech.2021-248
4568-4572
@inproceedings{kim21m_interspeech, title = {{PQK: Model Compression via Pruning, Quantization, and Knowledge Distillation}}, author = {Jangho Kim and Simyung Chang and Nojun Kwak}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4568--4572}, doi = {10.21437/Interspeech.2021-2...
As edge devices become prevalent, deploying Deep Neural Networks (DNN) on edge devices has become a critical issue. However, DNN requires a high computational resource which is rarely available for edge devices. To handle this, we propose a novel model compression method for the devices with limited computational resou...
2106.14681
title_snapshot
nagaraja21_interspeech
Collaborative Training of Acoustic Encoders for Speech Recognition
[ "Varun Nagaraja", "Yangyang Shi", "Ganesh Venkatesh", "Ozlem Kalinli", "Michael L. Seltzer", "Vikas Chandra" ]
https://www.isca-archive.org/interspeech_2021/nagaraja21_interspeech.html
https://www.isca-archive.org/interspeech_2021/nagaraja21_interspeech.pdf
10.21437/Interspeech.2021-354
4573-4577
@inproceedings{nagaraja21_interspeech, title = {{Collaborative Training of Acoustic Encoders for Speech Recognition}}, author = {Varun Nagaraja and Yangyang Shi and Ganesh Venkatesh and Ozlem Kalinli and Michael L. Seltzer and Vikas Chandra}, year = {2021}, booktitle = {{Interspeech 2021}}, pages ...
On-device speech recognition requires training models of different sizes for deploying on devices with various computational budgets. When building such different models, we can benefit from training them jointly to take advantage of the knowledge shared between them. Joint training is also efficient since it reduces t...
2106.08960
title_snapshot
wang21ha_interspeech
Efficient Conformer with Prob-Sparse Attention Mechanism for End-to-End Speech Recognition
[ "Xiong Wang", "Sining Sun", "Lei Xie", "Long Ma" ]
https://www.isca-archive.org/interspeech_2021/wang21ha_interspeech.html
https://www.isca-archive.org/interspeech_2021/wang21ha_interspeech.pdf
10.21437/Interspeech.2021-415
4578-4582
@inproceedings{wang21ha_interspeech, title = {{Efficient Conformer with Prob-Sparse Attention Mechanism for End-to-End Speech Recognition}}, author = {Xiong Wang and Sining Sun and Lei Xie and Long Ma}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4578--4582}, doi = {10.21...
End-to-end models are favored in automatic speech recognition (ASR) because of their simplified system structure and superior performance. Among these models, Transformer and Conformer have achieved state-of-the-art recognition accuracy in which self-attention plays a vital role in capturing important global informatio...
2106.09236
title_judge
parcollet21_interspeech
The Energy and Carbon Footprint of Training End-to-End Speech Recognizers
[ "Titouan Parcollet", "Mirco Ravanelli" ]
https://www.isca-archive.org/interspeech_2021/parcollet21_interspeech.html
https://www.isca-archive.org/interspeech_2021/parcollet21_interspeech.pdf
10.21437/Interspeech.2021-456
4583-4587
@inproceedings{parcollet21_interspeech, title = {{The Energy and Carbon Footprint of Training End-to-End Speech Recognizers}}, author = {Titouan Parcollet and Mirco Ravanelli}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4583--4587}, doi = {10.21437/Interspeech.2021-456},...
Deep learning contributes to reaching higher levels of artificial intelligence. Due to its pervasive adoption, however, growing concerns on the environmental impact of this technology have been raised. In particular, the energy consumed at training and inference time by modern neural networks is far from being negligib...
null
null
chen21v_interspeech
Graph-Based Label Propagation for Semi-Supervised Speaker Identification
[ "Long Chen", "Venkatesh Ravichandran", "Andreas Stolcke" ]
https://www.isca-archive.org/interspeech_2021/chen21v_interspeech.html
https://www.isca-archive.org/interspeech_2021/chen21v_interspeech.pdf
10.21437/Interspeech.2021-1209
4588-4592
@inproceedings{chen21v_interspeech, title = {{Graph-Based Label Propagation for Semi-Supervised Speaker Identification}}, author = {Long Chen and Venkatesh Ravichandran and Andreas Stolcke}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4588--4592}, doi = {10.21437/Interspe...
Speaker identification in the household scenario (e.g., for smart speakers) is typically based on only a few enrollment utterances but a much larger set of unlabeled data, suggesting semi-supervised learning to improve speaker profiles. We propose a graph-based semi-supervised learning approach for speaker identificati...
2106.08207
title_snapshot
li21q_interspeech
Fusion of Embeddings Networks for Robust Combination of Text Dependent and Independent Speaker Recognition
[ "Ruirui Li", "Chelsea J.-T. Ju", "Zeya Chen", "Hongda Mao", "Oguz Elibol", "Andreas Stolcke" ]
https://www.isca-archive.org/interspeech_2021/li21q_interspeech.html
https://www.isca-archive.org/interspeech_2021/li21q_interspeech.pdf
10.21437/Interspeech.2021-3
4593-4597
@inproceedings{li21q_interspeech, title = {{Fusion of Embeddings Networks for Robust Combination of Text Dependent and Independent Speaker Recognition}}, author = {Ruirui Li and Chelsea J.-T. Ju and Zeya Chen and Hongda Mao and Oguz Elibol and Andreas Stolcke}, year = {2021}, booktitle = {{Interspee...
By implicitly recognizing a user based on his/her speech input, speaker identification enables many downstream applications, such as personalized system behavior and expedited shopping checkouts. Based on whether the speech content is constrained or not, both text-dependent (TD) and text-independent (TI) speaker recogn...
2106.10169
title_snapshot
cumani21_interspeech
A Generative Model for Duration-Dependent Score Calibration
[ "Sandro Cumani", "Salvatore Sarni" ]
https://www.isca-archive.org/interspeech_2021/cumani21_interspeech.html
https://www.isca-archive.org/interspeech_2021/cumani21_interspeech.pdf
10.21437/Interspeech.2021-114
4598-4602
@inproceedings{cumani21_interspeech, title = {{A Generative Model for Duration-Dependent Score Calibration}}, author = {Sandro Cumani and Salvatore Sarni}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4598--4602}, doi = {10.21437/Interspeech.2021-114}, issn = {2958-...
In this work we introduce a generative score calibration model for speaker verification systems able to explicitly account for utterance-dependent miscalibration sources, with a focus on segment duration. The model is theoretically motivated by an analysis of the effects of distribution mismatch on the scores produced ...
null
null
pelecanos21_interspeech
Dr-Vectors: Decision Residual Networks and an Improved Loss for Speaker Recognition
[ "Jason Pelecanos", "Quan Wang", "Ignacio Lopez Moreno" ]
https://www.isca-archive.org/interspeech_2021/pelecanos21_interspeech.html
https://www.isca-archive.org/interspeech_2021/pelecanos21_interspeech.pdf
10.21437/Interspeech.2021-641
4603-4607
@inproceedings{pelecanos21_interspeech, title = {{Dr-Vectors: Decision Residual Networks and an Improved Loss for Speaker Recognition}}, author = {Jason Pelecanos and Quan Wang and Ignacio Lopez Moreno}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4603--4607}, doi = {10.2...
Many neural network speaker recognition systems model each speaker using a fixed-dimensional embedding vector. These embeddings are generally compared using either linear or 2nd-order scoring and, until recently, do not handle utterance-specific uncertainty. In this work we propose scoring these representations in a wa...
2104.01989
title_snapshot
kataria21b_interspeech
Multi-Channel Speaker Verification for Single and Multi-Talker Speech
[ "Saurabh Kataria", "Shi-Xiong Zhang", "Dong Yu" ]
https://www.isca-archive.org/interspeech_2021/kataria21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/kataria21b_interspeech.pdf
10.21437/Interspeech.2021-681
4608-4612
@inproceedings{kataria21b_interspeech, title = {{Multi-Channel Speaker Verification for Single and Multi-Talker Speech}}, author = {Saurabh Kataria and Shi-Xiong Zhang and Dong Yu}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4608--4612}, doi = {10.21437/Interspeech.2021-...
To improve speaker verification in real scenarios with interference speakers, noise, and reverberation, we propose to bring together advancements made in multi-channel speech features. Specifically, we combine spectral , spatial , and directional features, which includes inter-channel phase difference, multichannel sin...
2010.12692
title_snapshot
padfield21_interspeech
Chronological Self-Training for Real-Time Speaker Diarization
[ "Dirk Padfield", "Daniel J. Liebling" ]
https://www.isca-archive.org/interspeech_2021/padfield21_interspeech.html
https://www.isca-archive.org/interspeech_2021/padfield21_interspeech.pdf
10.21437/Interspeech.2021-822
4613-4617
@inproceedings{padfield21_interspeech, title = {{Chronological Self-Training for Real-Time Speaker Diarization}}, author = {Dirk Padfield and Daniel J. Liebling}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4613--4617}, doi = {10.21437/Interspeech.2021-822}, issn =...
Diarization partitions an audio stream into segments based on the voices of the speakers. Real-time diarization systems that include an enrollment step should limit enrollment training samples to reduce user interaction time. Although training on a small number of samples yields poor performance, we show that the accur...
2208.03393
title_snapshot
xiao21b_interspeech
Adaptive Margin Circle Loss for Speaker Verification
[ "Runqiu Xiao", "Xiaoxiao Miao", "Wenchao Wang", "Pengyuan Zhang", "Bin Cai", "Liuping Luo" ]
https://www.isca-archive.org/interspeech_2021/xiao21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/xiao21b_interspeech.pdf
10.21437/Interspeech.2021-1043
4618-4622
@inproceedings{xiao21b_interspeech, title = {{Adaptive Margin Circle Loss for Speaker Verification}}, author = {Runqiu Xiao and Xiaoxiao Miao and Wenchao Wang and Pengyuan Zhang and Bin Cai and Liuping Luo}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4618--4622}, doi = {...
Deep-Neural-Network (DNN) based speaker verification systems use the angular softmax loss with margin penalties to enhance the intra-class compactness of speaker embeddings, which achieved remarkable performance. In this paper, we propose a novel angular loss function called adaptive margin circle loss for speaker veri...
2106.08004
title_snapshot
obrien21b_interspeech
Presentation Matters: Evaluating Speaker Identification Tasks
[ "Benjamin O’Brien", "Christine Meunier", "Alain Ghio" ]
https://www.isca-archive.org/interspeech_2021/obrien21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/obrien21b_interspeech.pdf
10.21437/Interspeech.2021-1211
4623-4627
@inproceedings{obrien21b_interspeech, title = {{Presentation Matters: Evaluating Speaker Identification Tasks}}, author = {Benjamin O’Brien and Christine Meunier and Alain Ghio}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4623--4627}, doi = {10.21437/Interspeech.2021-121...
This paper details our evaluations and comparisons of speaker identification (SID) performance by listeners across different tasks. Experiment 1 participants completed traditional target-lineup (1-out-of-N speakers or out-of-set speaker) and binary (speaker verification) tasks. Experiment 2 participants completed trial...
null
null
tong21_interspeech
Automatic Error Correction for Speaker Embedding Learning with Noisy Labels
[ "Fuchuan Tong", "Yan Liu", "Song Li", "Jie Wang", "Lin Li", "Qingyang Hong" ]
https://www.isca-archive.org/interspeech_2021/tong21_interspeech.html
https://www.isca-archive.org/interspeech_2021/tong21_interspeech.pdf
10.21437/Interspeech.2021-2021
4628-4632
@inproceedings{tong21_interspeech, title = {{Automatic Error Correction for Speaker Embedding Learning with Noisy Labels}}, author = {Fuchuan Tong and Yan Liu and Song Li and Jie Wang and Lin Li and Qingyang Hong}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4628--4632}, doi ...
Despite the superior performance deep neural networks have achieved in speaker verification tasks, much of their success benefits from the availability of large-scale and carefully labeled datasets. However, noisy labels often occur during data collection. In this paper, we propose an automatic error correction method ...
null
null
liao21_interspeech
An Integrated Framework for Two-Pass Personalized Voice Trigger
[ "Dexin Liao", "Jing Li", "Yiming Zhi", "Song Li", "Qingyang Hong", "Lin Li" ]
https://www.isca-archive.org/interspeech_2021/liao21_interspeech.html
https://www.isca-archive.org/interspeech_2021/liao21_interspeech.pdf
10.21437/Interspeech.2021-2161
4633-4637
@inproceedings{liao21_interspeech, title = {{An Integrated Framework for Two-Pass Personalized Voice Trigger}}, author = {Dexin Liao and Jing Li and Yiming Zhi and Song Li and Qingyang Hong and Lin Li}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4633--4637}, doi = {10.21...
In this paper, we present the XMUSPEECH system for Task 1 of 2020 Personalized Voice Trigger Challenge (PVTC2020). Task 1 is a joint wake-up word detection with speaker verification on close talking data. The whole system consists of a keyword spotting (KWS) sub-system and a speaker verification (SV) sub-system. For th...
2106.15950
title_snapshot
lian21_interspeech
Masked Proxy Loss for Text-Independent Speaker Verification
[ "Jiachen Lian", "Aiswarya Vinod Kumar", "Hira Dhamyal", "Bhiksha Raj", "Rita Singh" ]
https://www.isca-archive.org/interspeech_2021/lian21_interspeech.html
https://www.isca-archive.org/interspeech_2021/lian21_interspeech.pdf
10.21437/Interspeech.2021-2190
4638-4642
@inproceedings{lian21_interspeech, title = {{Masked Proxy Loss for Text-Independent Speaker Verification}}, author = {Jiachen Lian and Aiswarya Vinod Kumar and Hira Dhamyal and Bhiksha Raj and Rita Singh}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4638--4642}, doi = {10...
Open-set speaker recognition can be regarded as a metric learning problem, which is to maximize inter-class variance and minimize intra-class variance. Supervised metric learning can be categorized into pair-based learning and proxy-based learning [1]. Most of the existing metric learning objectives belong to the forme...
2011.04491
title_snapshot
lee21h_interspeech
STYLER: Style Factor Modeling with Rapidity and Robustness via Speech Decomposition for Expressive and Controllable Neural Text to Speech
[ "Keon Lee", "Kyumin Park", "Daeyoung Kim" ]
https://www.isca-archive.org/interspeech_2021/lee21h_interspeech.html
https://www.isca-archive.org/interspeech_2021/lee21h_interspeech.pdf
10.21437/Interspeech.2021-838
4643-4647
@inproceedings{lee21h_interspeech, title = {{STYLER: Style Factor Modeling with Rapidity and Robustness via Speech Decomposition for Expressive and Controllable Neural Text to Speech}}, author = {Keon Lee and Kyumin Park and Daeyoung Kim}, year = {2021}, booktitle = {{Interspeech 2021}}, pages ...
Previous works on neural text-to-speech (TTS) have been addressed on limited speed in training and inference time, robustness for difficult synthesis conditions, expressiveness, and controllability. Although several approaches resolve some limitations, there has been no attempt to solve all weaknesses at once. In this ...
2103.09474
title_snapshot
liu21p_interspeech
Reinforcement Learning for Emotional Text-to-Speech Synthesis with Improved Emotion Discriminability
[ "Rui Liu", "Berrak Sisman", "Haizhou Li" ]
https://www.isca-archive.org/interspeech_2021/liu21p_interspeech.html
https://www.isca-archive.org/interspeech_2021/liu21p_interspeech.pdf
10.21437/Interspeech.2021-1236
4648-4652
@inproceedings{liu21p_interspeech, title = {{Reinforcement Learning for Emotional Text-to-Speech Synthesis with Improved Emotion Discriminability}}, author = {Rui Liu and Berrak Sisman and Haizhou Li}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4648--4652}, doi = {10.214...
Emotional text-to-speech synthesis (ETTS) has seen much progress in recent years. However, the generated voice is often not perceptually identifiable by its intended emotion category. To address this problem, we propose a new interactive training paradigm for ETTS, denoted as i-ETTS , which seeks to directly improve th...
2104.01408
title_snapshot
sivaprasad21_interspeech
Emotional Prosody Control for Speech Generation
[ "Sarath Sivaprasad", "Saiteja Kosgi", "Vineet Gandhi" ]
https://www.isca-archive.org/interspeech_2021/sivaprasad21_interspeech.html
https://www.isca-archive.org/interspeech_2021/sivaprasad21_interspeech.pdf
10.21437/Interspeech.2021-307
4653-4657
@inproceedings{sivaprasad21_interspeech, title = {{Emotional Prosody Control for Speech Generation}}, author = {Sarath Sivaprasad and Saiteja Kosgi and Vineet Gandhi}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4653--4657}, doi = {10.21437/Interspeech.2021-307}, issn ...
Machine-generated speech is characterized by its limited or unnatural emotional variation. Current text to speech systems generates speech with either a flat emotion, emotion selected from a predefined set, average variation learned from prosody sequences in training data or transferred from a source style. We propose ...
2111.04730
title_snapshot
cong21b_interspeech
Controllable Context-Aware Conversational Speech Synthesis
[ "Jian Cong", "Shan Yang", "Na Hu", "Guangzhi Li", "Lei Xie", "Dan Su" ]
https://www.isca-archive.org/interspeech_2021/cong21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/cong21b_interspeech.pdf
10.21437/Interspeech.2021-412
4658-4662
@inproceedings{cong21b_interspeech, title = {{Controllable Context-Aware Conversational Speech Synthesis}}, author = {Jian Cong and Shan Yang and Na Hu and Guangzhi Li and Lei Xie and Dan Su}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4658--4662}, doi = {10.21437/Inters...
In spoken conversations, spontaneous behaviors like filled pause and prolongations always happen. Conversational partner tends to align features of their speech with their interlocutor which is known as entrainment. To produce human-like conversations, we propose a unified controllable spontaneous conversational speech...
2106.10828
title_snapshot
kim21n_interspeech
Expressive Text-to-Speech Using Style Tag
[ "Minchan Kim", "Sung Jun Cheon", "Byoung Jin Choi", "Jong Jin Kim", "Nam Soo Kim" ]
https://www.isca-archive.org/interspeech_2021/kim21n_interspeech.html
https://www.isca-archive.org/interspeech_2021/kim21n_interspeech.pdf
10.21437/Interspeech.2021-465
4663-4667
@inproceedings{kim21n_interspeech, title = {{Expressive Text-to-Speech Using Style Tag}}, author = {Minchan Kim and Sung Jun Cheon and Byoung Jin Choi and Jong Jin Kim and Nam Soo Kim}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4663--4667}, doi = {10.21437/Interspeech.2...
As recent text-to-speech (TTS) systems have been rapidly improved in speech quality and generation speed, many researchers now focus on a more challenging issue: expressive TTS. To control speaking styles, existing expressive TTS models use categorical style index or reference speech as style input. In this work, we pr...
2104.00436
title_snapshot
yan21d_interspeech
Adaptive Text to Speech for Spontaneous Style
[ "Yuzi Yan", "Xu Tan", "Bohan Li", "Guangyan Zhang", "Tao Qin", "Sheng Zhao", "Yuan Shen", "Wei-Qiang Zhang", "Tie-Yan Liu" ]
https://www.isca-archive.org/interspeech_2021/yan21d_interspeech.html
https://www.isca-archive.org/interspeech_2021/yan21d_interspeech.pdf
10.21437/Interspeech.2021-584
4668-4672
@inproceedings{yan21d_interspeech, title = {{Adaptive Text to Speech for Spontaneous Style}}, author = {Yuzi Yan and Xu Tan and Bohan Li and Guangyan Zhang and Tao Qin and Sheng Zhao and Yuan Shen and Wei-Qiang Zhang and Tie-Yan Liu}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4...
While recent text to speech (TTS) models perform very well in synthesizing reading-style (e.g., audiobook) speech, it is still challenging to synthesize spontaneous-style speech (e.g., podcast or conversation), mainly because of two reasons: 1) the lack of training data for spontaneous speech; 2) the difficulty in mode...
2107.02530
title_judge
li21r_interspeech
Towards Multi-Scale Style Control for Expressive Speech Synthesis
[ "Xiang Li", "Changhe Song", "Jingbei Li", "Zhiyong Wu", "Jia Jia", "Helen Meng" ]
https://www.isca-archive.org/interspeech_2021/li21r_interspeech.html
https://www.isca-archive.org/interspeech_2021/li21r_interspeech.pdf
10.21437/Interspeech.2021-947
4673-4677
@inproceedings{li21r_interspeech, title = {{Towards Multi-Scale Style Control for Expressive Speech Synthesis}}, author = {Xiang Li and Changhe Song and Jingbei Li and Zhiyong Wu and Jia Jia and Helen Meng}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4673--4677}, doi = {...
This paper introduces a multi-scale speech style modeling method for end-to-end expressive speech synthesis. The proposed method employs a multi-scale reference encoder to extract both the global-scale utterance-level and the local-scale quasi-phoneme-level style features of the target speech, which are then fed into t...
2104.03521
title_snapshot
pan21d_interspeech
Cross-Speaker Style Transfer with Prosody Bottleneck in Neural Speech Synthesis
[ "Shifeng Pan", "Lei He" ]
https://www.isca-archive.org/interspeech_2021/pan21d_interspeech.html
https://www.isca-archive.org/interspeech_2021/pan21d_interspeech.pdf
10.21437/Interspeech.2021-979
4678-4682
@inproceedings{pan21d_interspeech, title = {{Cross-Speaker Style Transfer with Prosody Bottleneck in Neural Speech Synthesis}}, author = {Shifeng Pan and Lei He}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4678--4682}, doi = {10.21437/Interspeech.2021-979}, issn =...
Cross-speaker style transfer is crucial to the applications of multi-style and expressive speech synthesis at scale. It does not require the target speakers to be experts in expressing all styles and to collect corresponding recordings for model training. However, the performances of existing style transfer methods are...
2107.12562
title_snapshot
tan21_interspeech
Fine-Grained Style Modeling, Transfer and Prediction in Text-to-Speech Synthesis via Phone-Level Content-Style Disentanglement
[ "Daxin Tan", "Tan Lee" ]
https://www.isca-archive.org/interspeech_2021/tan21_interspeech.html
https://www.isca-archive.org/interspeech_2021/tan21_interspeech.pdf
10.21437/Interspeech.2021-1129
4683-4687
@inproceedings{tan21_interspeech, title = {{Fine-Grained Style Modeling, Transfer and Prediction in Text-to-Speech Synthesis via Phone-Level Content-Style Disentanglement}}, author = {Daxin Tan and Tan Lee}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4683--4687}, doi = {...
This paper presents a novel design of neural network system for fine-grained style modeling, transfer and prediction in expressive text-to-speech (TTS) synthesis. Fine-grained modeling is realized by extracting style embeddings from the mel-spectrograms of phone-level speech segments. Collaborative learning and adversa...
2011.03943
title_snapshot
an21b_interspeech
Improving Performance of Seen and Unseen Speech Style Transfer in End-to-End Neural TTS
[ "Xiaochun An", "Frank K. Soong", "Lei Xie" ]
https://www.isca-archive.org/interspeech_2021/an21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/an21b_interspeech.pdf
10.21437/Interspeech.2021-1407
4688-4692
@inproceedings{an21b_interspeech, title = {{Improving Performance of Seen and Unseen Speech Style Transfer in End-to-End Neural TTS}}, author = {Xiaochun An and Frank K. Soong and Lei Xie}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4688--4692}, doi = {10.21437/Interspee...
End-to-end neural TTS training has shown improved performance in speech style transfer. However, the improvement is still limited by the training data in both target styles and speakers. Inadequate style transfer performance occurs when the trained TTS tries to transfer the speech to a target style from a new speaker w...
2106.10003
title_snapshot
shechtman21_interspeech
Synthesis of Expressive Speaking Styles with Limited Training Data in a Multi-Speaker, Prosody-Controllable Sequence-to-Sequence Architecture
[ "Slava Shechtman", "Raul Fernandez", "Alexander Sorin", "David Haws" ]
https://www.isca-archive.org/interspeech_2021/shechtman21_interspeech.html
https://www.isca-archive.org/interspeech_2021/shechtman21_interspeech.pdf
10.21437/Interspeech.2021-1446
4693-4697
@inproceedings{shechtman21_interspeech, title = {{Synthesis of Expressive Speaking Styles with Limited Training Data in a Multi-Speaker, Prosody-Controllable Sequence-to-Sequence Architecture}}, author = {Slava Shechtman and Raul Fernandez and Alexander Sorin and David Haws}, year = {2021}, booktitl...
Although Sequence-to-Sequence (S2S) architectures have become state-of-the-art in speech synthesis, the best models benefit from access to moderate-to-large amounts of training data, posing a resource bottleneck when we are interested in generating speech in a variety of expressive styles. In this work we explore a S2S...
null
null
dao21_interspeech
Intent Detection and Slot Filling for Vietnamese
[ "Mai Hoang Dao", "Thinh Hung Truong", "Dat Quoc Nguyen" ]
https://www.isca-archive.org/interspeech_2021/dao21_interspeech.html
https://www.isca-archive.org/interspeech_2021/dao21_interspeech.pdf
10.21437/Interspeech.2021-618
4698-4702
@inproceedings{dao21_interspeech, title = {{Intent Detection and Slot Filling for Vietnamese}}, author = {Mai Hoang Dao and Thinh Hung Truong and Dat Quoc Nguyen}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4698--4702}, doi = {10.21437/Interspeech.2021-618}, issn ...
Intent detection and slot filling are important tasks in spoken and natural language understanding. However, Vietnamese is a low-resource language in these research topics. In this paper, we present the first public intent detection and slot filling dataset for Vietnamese. In addition, we also propose a joint model for...
2104.02021
title_snapshot
lin21k_interspeech
Augmenting Slot Values and Contexts for Spoken Language Understanding with Pretrained Models
[ "Haitao Lin", "Lu Xiang", "Yu Zhou", "Jiajun Zhang", "Chengqing Zong" ]
https://www.isca-archive.org/interspeech_2021/lin21k_interspeech.html
https://www.isca-archive.org/interspeech_2021/lin21k_interspeech.pdf
10.21437/Interspeech.2021-55
4703-4707
@inproceedings{lin21k_interspeech, title = {{Augmenting Slot Values and Contexts for Spoken Language Understanding with Pretrained Models}}, author = {Haitao Lin and Lu Xiang and Yu Zhou and Jiajun Zhang and Chengqing Zong}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4703--4707}...
Spoken Language Understanding (SLU) is one essential step in building a dialogue system. Due to the expensive cost of obtaining the labeled data, SLU suffers from the data scarcity problem. Therefore, in this paper, we focus on data augmentation for slot filling task in SLU. To achieve that, we aim at generating more d...
2108.08451
title_snapshot
gaspers21_interspeech
The Impact of Intent Distribution Mismatch on Semi-Supervised Spoken Language Understanding
[ "Judith Gaspers", "Quynh Do", "Daniil Sorokin", "Patrick Lehnen" ]
https://www.isca-archive.org/interspeech_2021/gaspers21_interspeech.html
https://www.isca-archive.org/interspeech_2021/gaspers21_interspeech.pdf
10.21437/Interspeech.2021-335
4708-4712
@inproceedings{gaspers21_interspeech, title = {{The Impact of Intent Distribution Mismatch on Semi-Supervised Spoken Language Understanding}}, author = {Judith Gaspers and Quynh Do and Daniil Sorokin and Patrick Lehnen}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4708--4712}, ...
With the expanding role of voice-controlled devices, bootstrapping spoken language understanding models from little labeled data becomes essential. Semi-supervised learning is a common technique to improve model performance when labeled data is scarce. In a real-world production system, the labeled data and the online ...
null
null
jiang21c_interspeech
Knowledge Distillation from BERT Transformer to Speech Transformer for Intent Classification
[ "Yidi Jiang", "Bidisha Sharma", "Maulik Madhavi", "Haizhou Li" ]
https://www.isca-archive.org/interspeech_2021/jiang21c_interspeech.html
https://www.isca-archive.org/interspeech_2021/jiang21c_interspeech.pdf
10.21437/Interspeech.2021-402
4713-4717
@inproceedings{jiang21c_interspeech, title = {{Knowledge Distillation from BERT Transformer to Speech Transformer for Intent Classification}}, author = {Yidi Jiang and Bidisha Sharma and Maulik Madhavi and Haizhou Li}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4713--4717}, do...
End-to-end intent classification using speech has numerous advantages compared to the conventional pipeline approach using automatic speech recognition (ASR), followed by natural language processing modules. It attempts to predict intent from speech without using an intermediate ASR module. However, such end-to-end fra...
2108.02598
title_snapshot
wang21ia_interspeech
Three-Module Modeling For End-to-End Spoken Language Understanding Using Pre-Trained DNN-HMM-Based Acoustic-Phonetic Model
[ "Nick J.C. Wang", "Lu Wang", "Yandan Sun", "Haimei Kang", "Dejun Zhang" ]
https://www.isca-archive.org/interspeech_2021/wang21ia_interspeech.html
https://www.isca-archive.org/interspeech_2021/wang21ia_interspeech.pdf
10.21437/Interspeech.2021-501
4718-4722
@inproceedings{wang21ia_interspeech, title = {{Three-Module Modeling For End-to-End Spoken Language Understanding Using Pre-Trained DNN-HMM-Based Acoustic-Phonetic Model}}, author = {Nick J.C. Wang and Lu Wang and Yandan Sun and Haimei Kang and Dejun Zhang}, year = {2021}, booktitle = {{Interspeech ...
In spoken language understanding (SLU), what the user says is converted to his/her intent. Recent work on end-to-end SLU has shown that accuracy can be improved via pre-training approaches. We revisit ideas presented by Lugosch et al. using speech pre-training and three-module modeling; however, to ease construction of...
2204.03315
title_snapshot
cha21_interspeech
Speak or Chat with Me: End-to-End Spoken Language Understanding System with Flexible Inputs
[ "Sujeong Cha", "Wangrui Hou", "Hyun Jung", "My Phung", "Michael Picheny", "Hong-Kwang J. Kuo", "Samuel Thomas", "Edmilson Morais" ]
https://www.isca-archive.org/interspeech_2021/cha21_interspeech.html
https://www.isca-archive.org/interspeech_2021/cha21_interspeech.pdf
10.21437/Interspeech.2021-788
4723-4727
@inproceedings{cha21_interspeech, title = {{Speak or Chat with Me: End-to-End Spoken Language Understanding System with Flexible Inputs}}, author = {Sujeong Cha and Wangrui Hou and Hyun Jung and My Phung and Michael Picheny and Hong-Kwang J. Kuo and Samuel Thomas and Edmilson Morais}, year = {2021}, ...
A major focus of recent research in spoken language understanding (SLU) has been on the end-to-end approach where a single model can predict intents directly from speech inputs without intermediate transcripts. However, this approach presents some challenges. First, since speech can be considered as personally identifi...
2104.05752
title_snapshot
zhang21ha_interspeech
End-to-End Cross-Lingual Spoken Language Understanding Model with Multilingual Pretraining
[ "Xianwei Zhang", "Liang He" ]
https://www.isca-archive.org/interspeech_2021/zhang21ha_interspeech.html
https://www.isca-archive.org/interspeech_2021/zhang21ha_interspeech.pdf
10.21437/Interspeech.2021-818
4728-4732
@inproceedings{zhang21ha_interspeech, title = {{End-to-End Cross-Lingual Spoken Language Understanding Model with Multilingual Pretraining}}, author = {Xianwei Zhang and Liang He}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4728--4732}, doi = {10.21437/Interspeech.2021-8...
The spoken language understanding (SLU) plays an essential role in the field of human-computer interaction. Most of the current SLU systems are cascade systems of automatic speech recognition (ASR) and natural language understanding (NLU). Error propagation and scarcity of annotated speech data are two common difficult...
null
null
saghir21_interspeech
Factorization-Aware Training of Transformers for Natural Language Understanding on the Edge
[ "Hamidreza Saghir", "Samridhi Choudhary", "Sepehr Eghbali", "Clement Chung" ]
https://www.isca-archive.org/interspeech_2021/saghir21_interspeech.html
https://www.isca-archive.org/interspeech_2021/saghir21_interspeech.pdf
10.21437/Interspeech.2021-1816
4733-4737
@inproceedings{saghir21_interspeech, title = {{Factorization-Aware Training of Transformers for Natural Language Understanding on the Edge}}, author = {Hamidreza Saghir and Samridhi Choudhary and Sepehr Eghbali and Clement Chung}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4733-...
Fine-tuning transformer-based models have shown to outperform other methods for many Natural Language Understanding (NLU) tasks. Recent studies to reduce the size of transformer models have achieved reductions of > 80%, making on-device inference on powerful devices possible. However, other resource-constrained devices...
null
null
saxon21_interspeech
End-to-End Spoken Language Understanding for Generalized Voice Assistants
[ "Michael Saxon", "Samridhi Choudhary", "Joseph P. McKenna", "Athanasios Mouchtaris" ]
https://www.isca-archive.org/interspeech_2021/saxon21_interspeech.html
https://www.isca-archive.org/interspeech_2021/saxon21_interspeech.pdf
10.21437/Interspeech.2021-1826
4738-4742
@inproceedings{saxon21_interspeech, title = {{End-to-End Spoken Language Understanding for Generalized Voice Assistants}}, author = {Michael Saxon and Samridhi Choudhary and Joseph P. McKenna and Athanasios Mouchtaris}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4738--4742}, d...
End-to-end (E2E) spoken language understanding (SLU) systems predict utterance semantics directly from speech using a single model. Previous work in this area has focused on targeted tasks in fixed domains, where the output semantic structure is assumed a priori and the input speech is of limited complexity. In this wo...
2106.09009
title_snapshot
han21f_interspeech
Bi-Directional Joint Neural Networks for Intent Classification and Slot Filling
[ "Soyeon Caren Han", "Siqu Long", "Huichun Li", "Henry Weld", "Josiah Poon" ]
https://www.isca-archive.org/interspeech_2021/han21f_interspeech.html
https://www.isca-archive.org/interspeech_2021/han21f_interspeech.pdf
10.21437/Interspeech.2021-2044
4743-4747
@inproceedings{han21f_interspeech, title = {{Bi-Directional Joint Neural Networks for Intent Classification and Slot Filling}}, author = {Soyeon Caren Han and Siqu Long and Huichun Li and Henry Weld and Josiah Poon}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4743--4747}, doi ...
Intent classification and slot filling are two critical tasks for natural language understanding. Traditionally the two tasks proceeded independently. However, more recently joint models for intent classification and slot filling have achieved state-of-the-art performance, and have proved that there exists a strong rel...
2202.13079
title_snapshot
cutler21_interspeech
INTERSPEECH 2021 Acoustic Echo Cancellation Challenge
[ "Ross Cutler", "Ando Saabas", "Tanel Parnamaa", "Markus Loide", "Sten Sootla", "Marju Purin", "Hannes Gamper", "Sebastian Braun", "Karsten Sorensen", "Robert Aichner", "Sriram Srinivasan" ]
https://www.isca-archive.org/interspeech_2021/cutler21_interspeech.html
https://www.isca-archive.org/interspeech_2021/cutler21_interspeech.pdf
10.21437/Interspeech.2021-1870
4748-4752
@inproceedings{cutler21_interspeech, title = {{INTERSPEECH 2021 Acoustic Echo Cancellation Challenge}}, author = {Ross Cutler and Ando Saabas and Tanel Parnamaa and Markus Loide and Sten Sootla and Marju Purin and Hannes Gamper and Sebastian Braun and Karsten Sorensen and Robert Aichner and Sriram Srinivasan...
The INTERSPEECH 2021 Acoustic Echo Cancellation Challenge is intended to stimulate research in the area of acoustic echo cancellation (AEC), which is an important part of speech enhancement and still a top issue in audio communication. Many recent AEC studies report good performance on synthetic datasets where the trai...
null
null
pfeifenberger21_interspeech
Acoustic Echo Cancellation with Cross-Domain Learning
[ "Lukas Pfeifenberger", "Matthias Zoehrer", "Franz Pernkopf" ]
https://www.isca-archive.org/interspeech_2021/pfeifenberger21_interspeech.html
https://www.isca-archive.org/interspeech_2021/pfeifenberger21_interspeech.pdf
10.21437/Interspeech.2021-85
4753-4757
@inproceedings{pfeifenberger21_interspeech, title = {{Acoustic Echo Cancellation with Cross-Domain Learning}}, author = {Lukas Pfeifenberger and Matthias Zoehrer and Franz Pernkopf}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4753--4757}, doi = {10.21437/Interspeech.2021...
This paper proposes the Cross-Domain Echo-Controller (CDEC), submitted to the Interspeech 2021 AEC-Challenge. The algorithm consists of three building blocks: (i) a Time-Delay Compensation (TDC) module, (ii) a frequency-domain block-based Acoustic Echo Canceler (AEC), and (iii) a Time-Domain Neural-Network (TD-NN) used...
null
null
zhang21ia_interspeech
F-T-LSTM Based Complex Network for Joint Acoustic Echo Cancellation and Speech Enhancement
[ "Shimin Zhang", "Yuxiang Kong", "Shubo Lv", "Yanxin Hu", "Lei Xie" ]
https://www.isca-archive.org/interspeech_2021/zhang21ia_interspeech.html
https://www.isca-archive.org/interspeech_2021/zhang21ia_interspeech.pdf
10.21437/Interspeech.2021-1359
4758-4762
@inproceedings{zhang21ia_interspeech, title = {{F-T-LSTM Based Complex Network for Joint Acoustic Echo Cancellation and Speech Enhancement}}, author = {Shimin Zhang and Yuxiang Kong and Shubo Lv and Yanxin Hu and Lei Xie}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4758--4762}, ...
With the increasing demand for audio communication and online conference, ensuring the robustness of Acoustic Echo Cancellation (AEC) under the complicated acoustic scenario including noise, reverberation and nonlinear distortion has become a top issue. Although there have been some traditional methods that consider no...
2106.07577
title_snapshot
seidel21_interspeech
Y-Net FCRN for Acoustic Echo and Noise Suppression
[ "Ernst Seidel", "Jan Franzen", "Maximilian Strake", "Tim Fingscheidt" ]
https://www.isca-archive.org/interspeech_2021/seidel21_interspeech.html
https://www.isca-archive.org/interspeech_2021/seidel21_interspeech.pdf
10.21437/Interspeech.2021-1590
4763-4767
@inproceedings{seidel21_interspeech, title = {{Y2-Net FCRN for Acoustic Echo and Noise Suppression}}, author = {Ernst Seidel and Jan Franzen and Maximilian Strake and Tim Fingscheidt}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4763--4767}, doi = {10.21437/Interspeech.20...
In recent years, deep neural networks (DNNs) were studied as an alternative to traditional acoustic echo cancellation (AEC) algorithms. The proposed models achieved remarkable performance for the separate tasks of AEC and residual echo suppression (RES). A promising network topology is a fully convolutional recurrent n...
2103.17189
title_judge
peng21f_interspeech
Acoustic Echo Cancellation Using Deep Complex Neural Network with Nonlinear Magnitude Compression and Phase Information
[ "Renhua Peng", "Linjuan Cheng", "Chengshi Zheng", "Xiaodong Li" ]
https://www.isca-archive.org/interspeech_2021/peng21f_interspeech.html
https://www.isca-archive.org/interspeech_2021/peng21f_interspeech.pdf
10.21437/Interspeech.2021-2022
4768-4772
@inproceedings{peng21f_interspeech, title = {{Acoustic Echo Cancellation Using Deep Complex Neural Network with Nonlinear Magnitude Compression and Phase Information}}, author = {Renhua Peng and Linjuan Cheng and Chengshi Zheng and Xiaodong Li}, year = {2021}, booktitle = {{Interspeech 2021}}, pag...
This paper describes a two-stage acoustic echo cancellation (AEC) and suppression framework for the INTERSPEECH2021 AEC Challenge. In the first stage, four parallel partitioned block frequency domain adaptive filters are used to cancel the linear echo components, where the far-end signal is delayed 0ms, 320ms, 640ms an...
null
null
ivry21_interspeech
Nonlinear Acoustic Echo Cancellation with Deep Learning
[ "Amir Ivry", "Israel Cohen", "Baruch Berdugo" ]
https://www.isca-archive.org/interspeech_2021/ivry21_interspeech.html
https://www.isca-archive.org/interspeech_2021/ivry21_interspeech.pdf
10.21437/Interspeech.2021-722
4773-4777
@inproceedings{ivry21_interspeech, title = {{Nonlinear Acoustic Echo Cancellation with Deep Learning}}, author = {Amir Ivry and Israel Cohen and Baruch Berdugo}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4773--4777}, doi = {10.21437/Interspeech.2021-722}, issn = ...
We propose a nonlinear acoustic echo cancellation system, which aims to model the echo path from the far-end signal to the near-end microphone in two parts. Inspired by the physical behavior of modern hands-free devices, we first introduce a novel neural network architecture that is specifically designed to model the n...
2106.13754
title_snapshot
green21_interspeech
Automatic Speech Recognition of Disordered Speech: Personalized Models Outperforming Human Listeners on Short Phrases
[ "Jordan R. Green", "Robert L. MacDonald", "Pan-Pan Jiang", "Julie Cattiau", "Rus Heywood", "Richard Cave", "Katie Seaver", "Marilyn A. Ladewig", "Jimmy Tobin", "Michael P. Brenner", "Philip C. Nelson", "Katrin Tomanek" ]
https://www.isca-archive.org/interspeech_2021/green21_interspeech.html
https://www.isca-archive.org/interspeech_2021/green21_interspeech.pdf
10.21437/Interspeech.2021-1384
4778-4782
@inproceedings{green21_interspeech, title = {{Automatic Speech Recognition of Disordered Speech: Personalized Models Outperforming Human Listeners on Short Phrases}}, author = {Jordan R. Green and Robert L. MacDonald and Pan-Pan Jiang and Julie Cattiau and Rus Heywood and Richard Cave and Katie Seaver and Ma...
This study evaluated the accuracy of personalized automatic speech recognition (ASR) for recognizing disordered speech from a large cohort of individuals with a wide range of underlying etiologies using an open vocabulary. The performance of these models was benchmarked relative to that of expert human transcribers and...
null
null
neumann21b_interspeech
Investigating the Utility of Multimodal Conversational Technology and Audiovisual Analytic Measures for the Assessment and Monitoring of Amyotrophic Lateral Sclerosis at Scale
[ "Michael Neumann", "Oliver Roesler", "Jackson Liscombe", "Hardik Kothare", "David Suendermann-Oeft", "David Pautler", "Indu Navar", "Aria Anvar", "Jochen Kumm", "Raquel Norel", "Ernest Fraenkel", "Alexander V. Sherman", "James D. Berry", "Gary L. Pattee", "Jun Wang", "Jordan R. Green",...
https://www.isca-archive.org/interspeech_2021/neumann21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/neumann21b_interspeech.pdf
10.21437/Interspeech.2021-1801
4783-4787
@inproceedings{neumann21b_interspeech, title = {{Investigating the Utility of Multimodal Conversational Technology and Audiovisual Analytic Measures for the Assessment and Monitoring of Amyotrophic Lateral Sclerosis at Scale}}, author = {Michael Neumann and Oliver Roesler and Jackson Liscombe and Hardik Koth...
We propose a cloud-based multimodal dialog platform for the remote assessment and monitoring of Amyotrophic Lateral Sclerosis (ALS) at scale. This paper presents our vision, technology setup, and an initial investigation of the efficacy of the various acoustic and visual speech metrics automatically extracted by the pl...
2104.07310
title_snapshot
hermann21_interspeech
Handling Acoustic Variation in Dysarthric Speech Recognition Systems Through Model Combination
[ "Enno Hermann", "Mathew Magimai-Doss" ]
https://www.isca-archive.org/interspeech_2021/hermann21_interspeech.html
https://www.isca-archive.org/interspeech_2021/hermann21_interspeech.pdf
10.21437/Interspeech.2021-2212
4788-4792
@inproceedings{hermann21_interspeech, title = {{Handling Acoustic Variation in Dysarthric Speech Recognition Systems Through Model Combination}}, author = {Enno Hermann and Mathew Magimai-Doss}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4788--4792}, doi = {10.21437/Inte...
Developing automatic speech recognition (ASR) systems that recognise dysarthric speech as well as control speech from unimpaired speakers remains challenging. Including more highly variable dysarthric speech during training can also negatively affect the performance on control speakers, which is not desirable when deve...
null
null
geng21b_interspeech
Spectro-Temporal Deep Features for Disordered Speech Assessment and Recognition
[ "Mengzhe Geng", "Shansong Liu", "Jianwei Yu", "Xurong Xie", "Shoukang Hu", "Zi Ye", "Zengrui Jin", "Xunying Liu", "Helen Meng" ]
https://www.isca-archive.org/interspeech_2021/geng21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/geng21b_interspeech.pdf
10.21437/Interspeech.2021-60
4793-4797
@inproceedings{geng21b_interspeech, title = {{Spectro-Temporal Deep Features for Disordered Speech Assessment and Recognition}}, author = {Mengzhe Geng and Shansong Liu and Jianwei Yu and Xurong Xie and Shoukang Hu and Zi Ye and Zengrui Jin and Xunying Liu and Helen Meng}, year = {2021}, booktitle =...
Automatic recognition of disordered speech remains a highly challenging task to date. Sources of variability commonly found in normal speech including accent, age or gender, when further compounded with the underlying causes of speech impairment and varying severity levels, create large diversity among speakers. To thi...
2201.05554
title_snapshot
gutz21_interspeech
Speaking with a KN95 Face Mask: ASR Performance and Speaker Compensation
[ "Sarah E. Gutz", "Hannah P. Rowe", "Jordan R. Green" ]
https://www.isca-archive.org/interspeech_2021/gutz21_interspeech.html
https://www.isca-archive.org/interspeech_2021/gutz21_interspeech.pdf
10.21437/Interspeech.2021-99
4798-4802
@inproceedings{gutz21_interspeech, title = {{Speaking with a KN95 Face Mask: ASR Performance and Speaker Compensation}}, author = {Sarah E. Gutz and Hannah P. Rowe and Jordan R. Green}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4798--4802}, doi = {10.21437/Interspeech.2...
The increasing prevalence of face masks in the United States due to the COVID-19 pandemic necessitates serious consideration of the functional impact of wearing a mask on speech. This study considers how the presence of a KN95 mask affects the performance of a commercial ASR system, Google Cloud Speech. We present evid...
null
null
jin21_interspeech
Adversarial Data Augmentation for Disordered Speech Recognition
[ "Zengrui Jin", "Mengzhe Geng", "Xurong Xie", "Jianwei Yu", "Shansong Liu", "Xunying Liu", "Helen Meng" ]
https://www.isca-archive.org/interspeech_2021/jin21_interspeech.html
https://www.isca-archive.org/interspeech_2021/jin21_interspeech.pdf
10.21437/Interspeech.2021-168
4803-4807
@inproceedings{jin21_interspeech, title = {{Adversarial Data Augmentation for Disordered Speech Recognition}}, author = {Zengrui Jin and Mengzhe Geng and Xurong Xie and Jianwei Yu and Shansong Liu and Xunying Liu and Helen Meng}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4803--...
Automatic recognition of disordered speech remains a highly challenging task to date. The underlying neuro-motor conditions, often compounded with co-occurring physical disabilities, lead to the difficulty in collecting large quantities of impaired speech required for ASR system development. To this end, data augmentat...
2108.00899
title_snapshot
xie21b_interspeech
Variational Auto-Encoder Based Variability Encoding for Dysarthric Speech Recognition
[ "Xurong Xie", "Rukiye Ruzi", "Xunying Liu", "Lan Wang" ]
https://www.isca-archive.org/interspeech_2021/xie21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/xie21b_interspeech.pdf
10.21437/Interspeech.2021-173
4808-4812
@inproceedings{xie21b_interspeech, title = {{Variational Auto-Encoder Based Variability Encoding for Dysarthric Speech Recognition}}, author = {Xurong Xie and Rukiye Ruzi and Xunying Liu and Lan Wang}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4808--4812}, doi = {10.214...
Dysarthric speech recognition is a challenging task due to acoustic variability and limited amount of available data. Diverse conditions of dysarthric speakers account for the acoustic variability, which make the variability difficult to be modeled precisely. This paper presents a variational auto-encoder based variabi...
2201.09422
title_snapshot
wang21ja_interspeech
Learning Explicit Prosody Models and Deep Speaker Embeddings for Atypical Voice Conversion
[ "Disong Wang", "Songxiang Liu", "Lifa Sun", "Xixin Wu", "Xunying Liu", "Helen Meng" ]
https://www.isca-archive.org/interspeech_2021/wang21ja_interspeech.html
https://www.isca-archive.org/interspeech_2021/wang21ja_interspeech.pdf
10.21437/Interspeech.2021-285
4813-4817
@inproceedings{wang21ja_interspeech, title = {{Learning Explicit Prosody Models and Deep Speaker Embeddings for Atypical Voice Conversion}}, author = {Disong Wang and Songxiang Liu and Lifa Sun and Xixin Wu and Xunying Liu and Helen Meng}, year = {2021}, booktitle = {{Interspeech 2021}}, pages ...
Though significant progress has been made for the voice conversion (VC) of typical speech, VC for atypical speech, e.g., dysarthric and second-language (L2) speech, remains a challenge, since it involves correcting for atypical prosody while maintaining speaker identity. To address this issue, we propose a VC system wi...
2011.01678
title_snapshot
deng21d_interspeech
Bayesian Parametric and Architectural Domain Adaptation of LF-MMI Trained TDNNs for Elderly and Dysarthric Speech Recognition
[ "Jiajun Deng", "Fabian Ritter Gutierrez", "Shoukang Hu", "Mengzhe Geng", "Xurong Xie", "Zi Ye", "Shansong Liu", "Jianwei Yu", "Xunying Liu", "Helen Meng" ]
https://www.isca-archive.org/interspeech_2021/deng21d_interspeech.html
https://www.isca-archive.org/interspeech_2021/deng21d_interspeech.pdf
10.21437/Interspeech.2021-289
4818-4822
@inproceedings{deng21d_interspeech, title = {{Bayesian Parametric and Architectural Domain Adaptation of LF-MMI Trained TDNNs for Elderly and Dysarthric Speech Recognition}}, author = {Jiajun Deng and Fabian Ritter Gutierrez and Shoukang Hu and Mengzhe Geng and Xurong Xie and Zi Ye and Shansong Liu and Jianw...
Automatic recognition of elderly and disordered speech remains a highly challenging task to date. Such data is not only difficult to collect in large quantities, but also exhibits a significant mismatch against normal speech trained ASR systems. To this end, conventional deep neural network model adaptation approaches ...
null
null
cai21c_interspeech
A Voice-Activated Switch for Persons with Motor and Speech Impairments: Isolated-Vowel Spotting Using Neural Networks
[ "Shanqing Cai", "Lisie Lillianfeld", "Katie Seaver", "Jordan R. Green", "Michael P. Brenner", "Philip C. Nelson", "D. Sculley" ]
https://www.isca-archive.org/interspeech_2021/cai21c_interspeech.html
https://www.isca-archive.org/interspeech_2021/cai21c_interspeech.pdf
10.21437/Interspeech.2021-330
4823-4827
@inproceedings{cai21c_interspeech, title = {{A Voice-Activated Switch for Persons with Motor and Speech Impairments: Isolated-Vowel Spotting Using Neural Networks}}, author = {Shanqing Cai and Lisie Lillianfeld and Katie Seaver and Jordan R. Green and Michael P. Brenner and Philip C. Nelson and D. Sculley}, ...
Severe speech impairments limit the precision and range of producible speech sounds. As a result, generic automatic speech recognition (ASR) and keyword spotting (KWS) systems fail to accurately recognize the utterances produced by individuals with severe speech impairments. This paper describes an approach in a simple...
null
null
chen21w_interspeech
Conformer Parrotron: A Faster and Stronger End-to-End Speech Conversion and Recognition Model for Atypical Speech
[ "Zhehuai Chen", "Bhuvana Ramabhadran", "Fadi Biadsy", "Xia Zhang", "Youzheng Chen", "Liyang Jiang", "Fang Chu", "Rohan Doshi", "Pedro J. Moreno" ]
https://www.isca-archive.org/interspeech_2021/chen21w_interspeech.html
https://www.isca-archive.org/interspeech_2021/chen21w_interspeech.pdf
10.21437/Interspeech.2021-676
4828-4832
@inproceedings{chen21w_interspeech, title = {{Conformer Parrotron: A Faster and Stronger End-to-End Speech Conversion and Recognition Model for Atypical Speech}}, author = {Zhehuai Chen and Bhuvana Ramabhadran and Fadi Biadsy and Xia Zhang and Youzheng Chen and Liyang Jiang and Fang Chu and Rohan Doshi and P...
Parrotron is an end-to-end personalizable model that enables many-to-one voice conversion (VC) and automated speech recognition (ASR) simultaneously for atypical speech. In this work, we present the next-generation Parrotron model with improvements in overall accuracy, training and inference speeds. The proposed archit...
null
null
macdonald21_interspeech
Disordered Speech Data Collection: Lessons Learned at 1 Million Utterances from Project Euphonia
[ "Robert L. MacDonald", "Pan-Pan Jiang", "Julie Cattiau", "Rus Heywood", "Richard Cave", "Katie Seaver", "Marilyn A. Ladewig", "Jimmy Tobin", "Michael P. Brenner", "Philip C. Nelson", "Jordan R. Green", "Katrin Tomanek" ]
https://www.isca-archive.org/interspeech_2021/macdonald21_interspeech.html
https://www.isca-archive.org/interspeech_2021/macdonald21_interspeech.pdf
10.21437/Interspeech.2021-697
4833-4837
@inproceedings{macdonald21_interspeech, title = {{Disordered Speech Data Collection: Lessons Learned at 1 Million Utterances from Project Euphonia}}, author = {Robert L. MacDonald and Pan-Pan Jiang and Julie Cattiau and Rus Heywood and Richard Cave and Katie Seaver and Marilyn A. Ladewig and Jimmy Tobin and ...
Speech samples from over 1000 individuals with impaired speech have been submitted for Project Euphonia, aimed at improving automated speech recognition systems for disordered speech. We provide an overview of the corpus, which recently passed 1 million utterances (>1300 hours), and review key lessons learned from this...
null
null
yeo21_interspeech
Automatic Severity Classification of Korean Dysarthric Speech Using Phoneme-Level Pronunciation Features
[ "Eun Jung Yeo", "Sunhee Kim", "Minhwa Chung" ]
https://www.isca-archive.org/interspeech_2021/yeo21_interspeech.html
https://www.isca-archive.org/interspeech_2021/yeo21_interspeech.pdf
10.21437/Interspeech.2021-1353
4838-4842
@inproceedings{yeo21_interspeech, title = {{Automatic Severity Classification of Korean Dysarthric Speech Using Phoneme-Level Pronunciation Features}}, author = {Eun Jung Yeo and Sunhee Kim and Minhwa Chung}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {4838--4842}, doi = ...
This paper proposes an automatic severity classification method for Korean dysarthric speech by using two types of phoneme-level pronunciation features. The first type is the percentage of correct phonemes, which consists of percentage of correct consonants, percentage of correct vowels, and percentage of total correct...
null
null
venugopalan21_interspeech
Comparing Supervised Models and Learned Speech Representations for Classifying Intelligibility of Disordered Speech on Selected Phrases
[ "Subhashini Venugopalan", "Joel Shor", "Manoj Plakal", "Jimmy Tobin", "Katrin Tomanek", "Jordan R. Green", "Michael P. Brenner" ]
https://www.isca-archive.org/interspeech_2021/venugopalan21_interspeech.html
https://www.isca-archive.org/interspeech_2021/venugopalan21_interspeech.pdf
10.21437/Interspeech.2021-1913
4843-4847
@inproceedings{venugopalan21_interspeech, title = {{Comparing Supervised Models and Learned Speech Representations for Classifying Intelligibility of Disordered Speech on Selected Phrases}}, author = {Subhashini Venugopalan and Joel Shor and Manoj Plakal and Jimmy Tobin and Katrin Tomanek and Jordan R. Green...
Automatic classification of disordered speech can provide an objective tool for identifying the presence and severity of a speech impairment. Classification approaches can also help identify hard-to-recognize speech samples to teach ASR systems about the variable manifestations of impaired speech. Here, we develop and ...
2107.03985
title_snapshot
mitra21_interspeech
Analysis and Tuning of a Voice Assistant System for Dysfluent Speech
[ "Vikramjit Mitra", "Zifang Huang", "Colin Lea", "Lauren Tooley", "Sarah Wu", "Darren Botten", "Ashwini Palekar", "Shrinath Thelapurath", "Panayiotis Georgiou", "Sachin Kajarekar", "Jefferey Bigham" ]
https://www.isca-archive.org/interspeech_2021/mitra21_interspeech.html
https://www.isca-archive.org/interspeech_2021/mitra21_interspeech.pdf
10.21437/Interspeech.2021-2006
4848-4852
@inproceedings{mitra21_interspeech, title = {{Analysis and Tuning of a Voice Assistant System for Dysfluent Speech}}, author = {Vikramjit Mitra and Zifang Huang and Colin Lea and Lauren Tooley and Sarah Wu and Darren Botten and Ashwini Palekar and Shrinath Thelapurath and Panayiotis Georgiou and Sachin Kajar...
Dysfluencies and variations in speech pronunciation can severely degrade speech recognition performance, and for many individuals with moderate-to-severe speech disorders, voice operated systems do not work. Current speech recognition systems are trained primarily with data from fluent speakers and as a consequence do ...
2106.11759
title_snapshot