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
kim21g_interspeech
Multi-Domain Knowledge Distillation via Uncertainty-Matching for End-to-End ASR Models
[ "Ho-Gyeong Kim", "Min-Joong Lee", "Hoshik Lee", "Tae Gyoon Kang", "Jihyun Lee", "Eunho Yang", "Sung Ju Hwang" ]
https://www.isca-archive.org/interspeech_2021/kim21g_interspeech.html
https://www.isca-archive.org/interspeech_2021/kim21g_interspeech.pdf
10.21437/Interspeech.2021-1169
2531-2535
@inproceedings{kim21g_interspeech, title = {{Multi-Domain Knowledge Distillation via Uncertainty-Matching for End-to-End ASR Models}}, author = {Ho-Gyeong Kim and Min-Joong Lee and Hoshik Lee and Tae Gyoon Kang and Jihyun Lee and Eunho Yang and Sung Ju Hwang}, year = {2021}, booktitle = {{Interspeec...
Knowledge Distillation basically matches predictive distributions of student and teacher networks to improve performance in an environment with model capacity and/or data constraints. However, it is well known that predictive distribution of neural networks not only tends to be overly confident, but also cannot directl...
null
null
macoskey21_interspeech
Learning a Neural Diff for Speech Models
[ "Jonathan Macoskey", "Grant P. Strimel", "Ariya Rastrow" ]
https://www.isca-archive.org/interspeech_2021/macoskey21_interspeech.html
https://www.isca-archive.org/interspeech_2021/macoskey21_interspeech.pdf
10.21437/Interspeech.2021-1575
2536-2540
@inproceedings{macoskey21_interspeech, title = {{Learning a Neural Diff for Speech Models}}, author = {Jonathan Macoskey and Grant P. Strimel and Ariya Rastrow}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2536--2540}, doi = {10.21437/Interspeech.2021-1575}, issn =...
As more speech processing applications execute locally on edge devices, a set of resource constraints must be considered. In this work we address one of these constraints, namely over-the-network data budgets for transferring models from server to device. We present neural update approaches for release of subsequent sp...
2108.01561
title_snapshot
zhang21p_interspeech
Stochastic Attention Head Removal: A Simple and Effective Method for Improving Transformer Based ASR Models
[ "Shucong Zhang", "Erfan Loweimi", "Peter Bell", "Steve Renals" ]
https://www.isca-archive.org/interspeech_2021/zhang21p_interspeech.html
https://www.isca-archive.org/interspeech_2021/zhang21p_interspeech.pdf
10.21437/Interspeech.2021-280
2541-2545
@inproceedings{zhang21p_interspeech, title = {{Stochastic Attention Head Removal: A Simple and Effective Method for Improving Transformer Based ASR Models}}, author = {Shucong Zhang and Erfan Loweimi and Peter Bell and Steve Renals}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {25...
Recently, Transformer based models have shown competitive automatic speech recognition (ASR) performance. One key factor in the success of these models is the multi-head attention mechanism. However, for trained models, we have previously observed that many attention matrices are close to diagonal, indicating the redun...
2011.04004
title_snapshot
xue21b_interspeech
Model-Agnostic Fast Adaptive Multi-Objective Balancing Algorithm for Multilingual Automatic Speech Recognition Model Training
[ "Jiabin Xue", "Tieran Zheng", "Jiqing Han" ]
https://www.isca-archive.org/interspeech_2021/xue21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/xue21b_interspeech.pdf
10.21437/Interspeech.2021-355
2546-2550
@inproceedings{xue21b_interspeech, title = {{Model-Agnostic Fast Adaptive Multi-Objective Balancing Algorithm for Multilingual Automatic Speech Recognition Model Training}}, author = {Jiabin Xue and Tieran Zheng and Jiqing Han}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2546--2...
This paper regards multilingual automatic speech recognition model training as a multi-objective problem because learning different languages may conflict, necessitating a trade-off. Most previous works on multilingual ASR model training mainly used data sampling to balance the performance of multiple languages but ign...
null
null
chang21b_interspeech
Towards Lifelong Learning of End-to-End ASR
[ "Heng-Jui Chang", "Hung-yi Lee", "Lin-shan Lee" ]
https://www.isca-archive.org/interspeech_2021/chang21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/chang21b_interspeech.pdf
10.21437/Interspeech.2021-563
2551-2555
@inproceedings{chang21b_interspeech, title = {{Towards Lifelong Learning of End-to-End ASR}}, author = {Heng-Jui Chang and Hung-yi Lee and Lin-shan Lee}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2551--2555}, doi = {10.21437/Interspeech.2021-563}, issn = {2958-17...
Automatic speech recognition (ASR) technologies today are primarily optimized for given datasets; thus, any changes in the application environment (e.g., acoustic conditions or topic domains) may inevitably degrade the performance. We can collect new data describing the new environment and fine-tune the system, but thi...
2104.01616
title_snapshot
leal21_interspeech
Self-Adaptive Distillation for Multilingual Speech Recognition: Leveraging Student Independence
[ "Isabel Leal", "Neeraj Gaur", "Parisa Haghani", "Brian Farris", "Pedro J. Moreno", "Manasa Prasad", "Bhuvana Ramabhadran", "Yun Zhu" ]
https://www.isca-archive.org/interspeech_2021/leal21_interspeech.html
https://www.isca-archive.org/interspeech_2021/leal21_interspeech.pdf
10.21437/Interspeech.2021-614
2556-2560
@inproceedings{leal21_interspeech, title = {{Self-Adaptive Distillation for Multilingual Speech Recognition: Leveraging Student Independence}}, author = {Isabel Leal and Neeraj Gaur and Parisa Haghani and Brian Farris and Pedro J. Moreno and Manasa Prasad and Bhuvana Ramabhadran and Yun Zhu}, year = {...
With a large population of the world speaking more than one language, multilingual automatic speech recognition (ASR) has gained popularity in the recent years. While lower resource languages can benefit from quality improvements in a multilingual ASR system, including unrelated or higher resource languages in the mix ...
null
null
xu21f_interspeech
Regularizing Word Segmentation by Creating Misspellings
[ "Hainan Xu", "Kartik Audhkhasi", "Yinghui Huang", "Jesse Emond", "Bhuvana Ramabhadran" ]
https://www.isca-archive.org/interspeech_2021/xu21f_interspeech.html
https://www.isca-archive.org/interspeech_2021/xu21f_interspeech.pdf
10.21437/Interspeech.2021-648
2561-2565
@inproceedings{xu21f_interspeech, title = {{Regularizing Word Segmentation by Creating Misspellings}}, author = {Hainan Xu and Kartik Audhkhasi and Yinghui Huang and Jesse Emond and Bhuvana Ramabhadran}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2561--2565}, doi = {10.2...
This work focuses on improving subword segmentation algorithms for end-to-end speech recognition models, and makes two major contributions. Firstly, we propose a novel word segmentation algorithm. The algorithm uses the same vocabulary generated by a regular wordpiece model, is easily extensible and supports a variety ...
null
null
wang21t_interspeech
Multitask Training with Text Data for End-to-End Speech Recognition
[ "Peidong Wang", "Tara N. Sainath", "Ron J. Weiss" ]
https://www.isca-archive.org/interspeech_2021/wang21t_interspeech.html
https://www.isca-archive.org/interspeech_2021/wang21t_interspeech.pdf
10.21437/Interspeech.2021-683
2566-2570
@inproceedings{wang21t_interspeech, title = {{Multitask Training with Text Data for End-to-End Speech Recognition}}, author = {Peidong Wang and Tara N. Sainath and Ron J. Weiss}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2566--2570}, doi = {10.21437/Interspeech.2021-683...
We propose a multitask training method for attention-based end-to-end speech recognition models. We regularize the decoder in a listen, attend, and spell model by multitask training it on both audio-text and text-only data. Trained on the 100-hour subset of LibriSpeech, the proposed method, without requiring an additio...
2010.14318
title_snapshot
chen21j_interspeech
Emitting Word Timings with HMM-Free End-to-End System in Automatic Speech Recognition
[ "Xianzhao Chen", "Hao Ni", "Yi He", "Kang Wang", "Zejun Ma", "Zongxia Xie" ]
https://www.isca-archive.org/interspeech_2021/chen21j_interspeech.html
https://www.isca-archive.org/interspeech_2021/chen21j_interspeech.pdf
10.21437/Interspeech.2021-894
2571-2575
@inproceedings{chen21j_interspeech, title = {{Emitting Word Timings with HMM-Free End-to-End System in Automatic Speech Recognition}}, author = {Xianzhao Chen and Hao Ni and Yi He and Kang Wang and Zejun Ma and Zongxia Xie}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2571--2575}...
Word timings, which mark the start and end times of each word in ASR results, play an important part in many applications, such as computer assisted language learning. To date, end-to-end (E2E) systems outperform conventional DNN-HMM hybrid systems in ASR accuracy but have challenges to obtain accurate word timings. In...
null
null
droppo21_interspeech
Scaling Laws for Acoustic Models
[ "Jasha Droppo", "Oguz Elibol" ]
https://www.isca-archive.org/interspeech_2021/droppo21_interspeech.html
https://www.isca-archive.org/interspeech_2021/droppo21_interspeech.pdf
10.21437/Interspeech.2021-1644
2576-2580
@inproceedings{droppo21_interspeech, title = {{Scaling Laws for Acoustic Models}}, author = {Jasha Droppo and Oguz Elibol}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2576--2580}, doi = {10.21437/Interspeech.2021-1644}, issn = {2958-1796}, }
There is a recent trend in machine learning to increase model quality by growing models to sizes previously thought to be unreasonable. Recent work has shown that autoregressive generative models with cross-entropy objective functions exhibit smooth power-law relationships, or scaling laws, that predict model quality f...
2106.09488
title_snapshot
billa21_interspeech
Leveraging Non-Target Language Resources to Improve ASR Performance in a Target Language
[ "Jayadev Billa" ]
https://www.isca-archive.org/interspeech_2021/billa21_interspeech.html
https://www.isca-archive.org/interspeech_2021/billa21_interspeech.pdf
10.21437/Interspeech.2021-1657
2581-2585
@inproceedings{billa21_interspeech, title = {{Leveraging Non-Target Language Resources to Improve ASR Performance in a Target Language}}, author = {Jayadev Billa}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2581--2585}, doi = {10.21437/Interspeech.2021-1657}, issn ...
This paper investigates approaches to improving automatic speech recognition (ASR) performance in a target language using resources in other languages. In particular, we assume that we have untranscribed speech in a different language and a well trained ASR system in yet another language. Concretely, we structure this ...
null
null
fasoli21_interspeech
4-Bit Quantization of LSTM-Based Speech Recognition Models
[ "Andrea Fasoli", "Chia-Yu Chen", "Mauricio Serrano", "Xiao Sun", "Naigang Wang", "Swagath Venkataramani", "George Saon", "Xiaodong Cui", "Brian Kingsbury", "Wei Zhang", "Zoltán Tüske", "Kailash Gopalakrishnan" ]
https://www.isca-archive.org/interspeech_2021/fasoli21_interspeech.html
https://www.isca-archive.org/interspeech_2021/fasoli21_interspeech.pdf
10.21437/Interspeech.2021-1962
2586-2590
@inproceedings{fasoli21_interspeech, title = {{4-Bit Quantization of LSTM-Based Speech Recognition Models}}, author = {Andrea Fasoli and Chia-Yu Chen and Mauricio Serrano and Xiao Sun and Naigang Wang and Swagath Venkataramani and George Saon and Xiaodong Cui and Brian Kingsbury and Wei Zhang and Zoltán Tüsk...
We investigate the impact of aggressive low-precision representations of weights and activations in two families of large LSTM-based architectures for Automatic Speech Recognition (ASR): hybrid Deep Bidirectional LSTM - Hidden Markov Models (DBLSTM-HMMs) and Recurrent Neural Network - Transducers (RNN-Ts). Using a 4-bi...
2108.12074
title_snapshot
masumura21_interspeech
Unified Autoregressive Modeling for Joint End-to-End Multi-Talker Overlapped Speech Recognition and Speaker Attribute Estimation
[ "Ryo Masumura", "Daiki Okamura", "Naoki Makishima", "Mana Ihori", "Akihiko Takashima", "Tomohiro Tanaka", "Shota Orihashi" ]
https://www.isca-archive.org/interspeech_2021/masumura21_interspeech.html
https://www.isca-archive.org/interspeech_2021/masumura21_interspeech.pdf
10.21437/Interspeech.2021-2043
2591-2595
@inproceedings{masumura21_interspeech, title = {{Unified Autoregressive Modeling for Joint End-to-End Multi-Talker Overlapped Speech Recognition and Speaker Attribute Estimation}}, author = {Ryo Masumura and Daiki Okamura and Naoki Makishima and Mana Ihori and Akihiko Takashima and Tomohiro Tanaka and Shota ...
In this paper, we present a novel modeling method for single-channel multi-talker overlapped automatic speech recognition (ASR) systems. Fully neural network based end-to-end models have dramatically improved the performance of multi-taker overlapped ASR tasks. One promising approach for end-to-end modeling is autoregr...
2107.01549
title_snapshot
meng21_interspeech
Minimum Word Error Rate Training with Language Model Fusion for End-to-End Speech Recognition
[ "Zhong Meng", "Yu Wu", "Naoyuki Kanda", "Liang Lu", "Xie Chen", "Guoli Ye", "Eric Sun", "Jinyu Li", "Yifan Gong" ]
https://www.isca-archive.org/interspeech_2021/meng21_interspeech.html
https://www.isca-archive.org/interspeech_2021/meng21_interspeech.pdf
10.21437/Interspeech.2021-2075
2596-2600
@inproceedings{meng21_interspeech, title = {{Minimum Word Error Rate Training with Language Model Fusion for End-to-End Speech Recognition}}, author = {Zhong Meng and Yu Wu and Naoyuki Kanda and Liang Lu and Xie Chen and Guoli Ye and Eric Sun and Jinyu Li and Yifan Gong}, year = {2021}, booktitle = ...
Integrating external language models (LMs) into end-to-end (E2E) models remains a challenging task for domain-adaptive speech recognition. Recently, internal language model estimation (ILME)-based LM fusion has shown significant word error rate (WER) reduction from Shallow Fusion by subtracting a weighted internal LM s...
2106.02302
title_snapshot
jiang21b_interspeech
Variable Frame Rate Acoustic Models Using Minimum Error Reinforcement Learning
[ "Dongcheng Jiang", "Chao Zhang", "Philip C. Woodland" ]
https://www.isca-archive.org/interspeech_2021/jiang21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/jiang21b_interspeech.pdf
10.21437/Interspeech.2021-2198
2601-2605
@inproceedings{jiang21b_interspeech, title = {{Variable Frame Rate Acoustic Models Using Minimum Error Reinforcement Learning}}, author = {Dongcheng Jiang and Chao Zhang and Philip C. Woodland}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2601--2605}, doi = {10.21437/Inte...
Frame selection in automatic speech recognition (ASR) systems can potentially improve the trade-off between speed and accuracy relative to fixed low frame rate methods. In this paper, a sequence training approach based on minimum error and reinforcement learning is proposed for a hybrid ASR system to operate at a varia...
null
null
kaland21_interspeech
How f0 and Phrase Position Affect Papuan Malay Word Identification
[ "Constantijn Kaland", "Matthew Gordon" ]
https://www.isca-archive.org/interspeech_2021/kaland21_interspeech.html
https://www.isca-archive.org/interspeech_2021/kaland21_interspeech.pdf
10.21437/Interspeech.2021-6
2606-2610
@inproceedings{kaland21_interspeech, title = {{How f0 and Phrase Position Affect Papuan Malay Word Identification}}, author = {Constantijn Kaland and Matthew Gordon}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2606--2610}, doi = {10.21437/Interspeech.2021-6}, issn ...
This paper reports a perception experiment on Papuan Malay, an Eastern Indonesian language for which phrase prosody is largely underresearched. While phrase-final f0 movements are the most prominent ones in this language, it remains to be seen to what extent they signal phrase boundaries (demarcating) or whether they c...
null
null
jespersen21_interspeech
On the Feasibility of the Danish Model of Intonational Transcription: Phonetic Evidence from Jutlandic Danish
[ "Anna Bothe Jespersen", "Pavel Šturm", "Míša Hejná" ]
https://www.isca-archive.org/interspeech_2021/jespersen21_interspeech.html
https://www.isca-archive.org/interspeech_2021/jespersen21_interspeech.pdf
10.21437/Interspeech.2021-190
2611-2615
@inproceedings{jespersen21_interspeech, title = {{On the Feasibility of the Danish Model of Intonational Transcription: Phonetic Evidence from Jutlandic Danish}}, author = {Anna Bothe Jespersen and Pavel Šturm and Míša Hejná}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2611--261...
Most of our knowledge of Danish f0 variation and intonation is based on the work of Grønnum and colleagues, who developed an a-phonological model in which a series of repeated “default” contours are superpositioned onto an overarching f0 slope. The current paper tests a range of predictions stemming from this model, mo...
null
null
meli21_interspeech
An Experiment in Paratone Detection in a Prosodically Annotated EAP Spoken Corpus
[ "Adrien Méli", "Nicolas Ballier", "Achille Falaise", "Alice Henderson" ]
https://www.isca-archive.org/interspeech_2021/meli21_interspeech.html
https://www.isca-archive.org/interspeech_2021/meli21_interspeech.pdf
10.21437/Interspeech.2021-294
2616-2620
@inproceedings{meli21_interspeech, title = {{An Experiment in Paratone Detection in a Prosodically Annotated EAP Spoken Corpus}}, author = {Adrien Méli and Nicolas Ballier and Achille Falaise and Alice Henderson}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2616--2620}, doi ...
This article describes an experiment in paratone detection based on a spoken corpus of English for Academic Purposes (EAP) recently automatically re-annotated with prosodic information. The Momel and INTSINT annotations were carried out using SPPAS. The EIIDA corpus was chosen as it offered long uninterrupted stretches...
null
null
gerazov21_interspeech
ProsoBeast Prosody Annotation Tool
[ "Branislav Gerazov", "Michael Wagner" ]
https://www.isca-archive.org/interspeech_2021/gerazov21_interspeech.html
https://www.isca-archive.org/interspeech_2021/gerazov21_interspeech.pdf
10.21437/Interspeech.2021-304
2621-2625
@inproceedings{gerazov21_interspeech, title = {{ProsoBeast Prosody Annotation Tool}}, author = {Branislav Gerazov and Michael Wagner}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2621--2625}, doi = {10.21437/Interspeech.2021-304}, issn = {2958-1796}, }
The labelling of speech corpora is a laborious and time-consuming process. The ProsoBeast Annotation Tool seeks to ease and accelerate this process by providing an interactive 2D representation of the prosodic landscape of the data, in which contours are distributed based on their similarity. This interactive map allow...
2104.02397
title_snapshot
tran21_interspeech
Assessing the Use of Prosody in Constituency Parsing of Imperfect Transcripts
[ "Trang Tran", "Mari Ostendorf" ]
https://www.isca-archive.org/interspeech_2021/tran21_interspeech.html
https://www.isca-archive.org/interspeech_2021/tran21_interspeech.pdf
10.21437/Interspeech.2021-373
2626-2630
@inproceedings{tran21_interspeech, title = {{Assessing the Use of Prosody in Constituency Parsing of Imperfect Transcripts}}, author = {Trang Tran and Mari Ostendorf}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2626--2630}, doi = {10.21437/Interspeech.2021-373}, issn ...
This work explores constituency parsing on automatically recognized transcripts of conversational speech. The neural parser is based on a sentence encoder that leverages word vectors contextualized with prosodic features, jointly learning prosodic feature extraction with parsing. We assess the utility of the prosody in...
2106.07794
title_snapshot
liu21i_interspeech
Targeted and Targetless Neutral Tones in Taiwanese Southern Min
[ "Roger Cheng-yen Liu", "Feng-fan Hsieh", "Yueh-chin Chang" ]
https://www.isca-archive.org/interspeech_2021/liu21i_interspeech.html
https://www.isca-archive.org/interspeech_2021/liu21i_interspeech.pdf
10.21437/Interspeech.2021-434
2631-2635
@inproceedings{liu21i_interspeech, title = {{Targeted and Targetless Neutral Tones in Taiwanese Southern Min}}, author = {Roger Cheng-yen Liu and Feng-fan Hsieh and Yueh-chin Chang}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2631--2635}, doi = {10.21437/Interspeech.2021...
This article is an acoustic study on the two types of neutral tone in Taiwanese Southern Min (TSM). Recording materials included a set of verb-clitic constructions with different preceding tones and clitics. Pitch contours in different conditions were compared using Smoothing Spline ANOVA. Our results confirmed that Ty...
null
null
gosy21_interspeech
The Interaction of Word Complexity and Word Duration in an Agglutinative Language
[ "Mária Gósy", "Kálmán Abari" ]
https://www.isca-archive.org/interspeech_2021/gosy21_interspeech.html
https://www.isca-archive.org/interspeech_2021/gosy21_interspeech.pdf
10.21437/Interspeech.2021-594
2636-2640
@inproceedings{gosy21_interspeech, title = {{The Interaction of Word Complexity and Word Duration in an Agglutinative Language}}, author = {Mária Gósy and Kálmán Abari}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2636--2640}, doi = {10.21437/Interspeech.2021-594}, issn...
The mental lexicon comprises the representations of various words either in a morphologically decomposed form, or in a conceptually non-decomposed form. The durations of mono-morphemic and multimorphemic words are assumed to contain information on the routes of their lexical access. The durations of Hungarian nouns wit...
null
null
pan21b_interspeech
Taiwan Min Nan (Taiwanese) Checked Tones Sound Change
[ "Ho-hsien Pan", "Shao-ren Lyu" ]
https://www.isca-archive.org/interspeech_2021/pan21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/pan21b_interspeech.pdf
10.21437/Interspeech.2021-672
2641-2645
@inproceedings{pan21b_interspeech, title = {{Taiwan Min Nan (Taiwanese) Checked Tones Sound Change}}, author = {Ho-hsien Pan and Shao-ren Lyu}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2641--2645}, doi = {10.21437/Interspeech.2021-672}, issn = {2958-1796}, }
The multifaced changes of Taiwan Min Nan (TMN) checked sandhi tones, S3 and S5 were investigated as well as the checked base tones, B3 and B5. Simultaneous EGG data, CQ_H and acoustic data, including duration, f0 offset at 80% vowel interval, and spectral tilt H1 * -A3 * from forty male and female speakers above 40 and...
null
null
jakob21_interspeech
In-Group Advantage in the Perception of Emotions: Evidence from Three Varieties of German
[ "Moritz Jakob", "Bettina Braun", "Katharina Zahner-Ritter" ]
https://www.isca-archive.org/interspeech_2021/jakob21_interspeech.html
https://www.isca-archive.org/interspeech_2021/jakob21_interspeech.pdf
10.21437/Interspeech.2021-1172
2646-2650
@inproceedings{jakob21_interspeech, title = {{In-Group Advantage in the Perception of Emotions: Evidence from Three Varieties of German}}, author = {Moritz Jakob and Bettina Braun and Katharina Zahner-Ritter}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2646--2650}, doi =...
Various studies on the perception of vocally expressed emotions have shown that recognition rates are higher if speaker and listener belong to the same cultural or linguistic group. This so-called in-group advantage is commonly attributed to prosodic differences in the expression of emotion across groups. Evidence come...
null
null
gobl21_interspeech
The LF Model in the Frequency Domain for Glottal Airflow Modelling Without Aliasing Distortion
[ "Christer Gobl" ]
https://www.isca-archive.org/interspeech_2021/gobl21_interspeech.html
https://www.isca-archive.org/interspeech_2021/gobl21_interspeech.pdf
10.21437/Interspeech.2021-1625
2651-2655
@inproceedings{gobl21_interspeech, title = {{The LF Model in the Frequency Domain for Glottal Airflow Modelling Without Aliasing Distortion}}, author = {Christer Gobl}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2651--2655}, doi = {10.21437/Interspeech.2021-1625}, issn...
Many of the commonly used voice source models are based on piecewise elementary functions defined in the time domain. The discrete-time implementation of such models generally causes aliasing distortion, which make them less useful for certain applications. This paper presents a method which eliminates this distortion....
null
null
wagner21_interspeech
Parsing Speech for Grouping and Prominence, and the Typology of Rhythm
[ "Michael Wagner", "Alvaro Iturralde Zurita", "Sijia Zhang" ]
https://www.isca-archive.org/interspeech_2021/wagner21_interspeech.html
https://www.isca-archive.org/interspeech_2021/wagner21_interspeech.pdf
10.21437/Interspeech.2021-1684
2656-2660
@inproceedings{wagner21_interspeech, title = {{Parsing Speech for Grouping and Prominence, and the Typology of Rhythm}}, author = {Michael Wagner and Alvaro Iturralde Zurita and Sijia Zhang}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2656--2660}, doi = {10.21437/Intersp...
Humans appear to be wired to perceive acoustic events rhythmically. English speakers, for example, tend to perceive alternating short and long sounds as a series of binary groups with a final beat (iambs), and alternating soft and loud sounds as a series of trochees. This generalization, often called the ‘Iambic-trocha...
null
null
mumtaz21_interspeech
Prosody of Case Markers in Urdu
[ "Benazir Mumtaz", "Massimiliano Canzi", "Miriam Butt" ]
https://www.isca-archive.org/interspeech_2021/mumtaz21_interspeech.html
https://www.isca-archive.org/interspeech_2021/mumtaz21_interspeech.pdf
10.21437/Interspeech.2021-1776
2661-2665
@inproceedings{mumtaz21_interspeech, title = {{Prosody of Case Markers in Urdu}}, author = {Benazir Mumtaz and Massimiliano Canzi and Miriam Butt}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2661--2665}, doi = {10.21437/Interspeech.2021-1776}, issn = {2958-1796}, ...
This paper studies the prosody of case clitics in Urdu, for which various different claims exist in the literature. We conducted a production experiment and controlled for effects potentially arising from the phonetics of the case clitics, the syntactic function they express and clausal position. We find that case clit...
null
null
stefansdottir21_interspeech
Articulatory Characteristics of Icelandic Voiced Fricative Lenition: Gradience, Categoricity, and Speaker/Gesture-Specific Effects
[ "Brynhildur Stefansdottir", "Francesco Burroni", "Sam Tilsen" ]
https://www.isca-archive.org/interspeech_2021/stefansdottir21_interspeech.html
https://www.isca-archive.org/interspeech_2021/stefansdottir21_interspeech.pdf
10.21437/Interspeech.2021-1903
2666-2670
@inproceedings{stefansdottir21_interspeech, title = {{Articulatory Characteristics of Icelandic Voiced Fricative Lenition: Gradience, Categoricity, and Speaker/Gesture-Specific Effects}}, author = {Brynhildur Stefansdottir and Francesco Burroni and Sam Tilsen}, year = {2021}, booktitle = {{Interspee...
Icelandic voiced fricatives frequently reduce in connected speech. However, systematic investigations of the phenomenon from acoustic and articulatory perspectives are lacking. To further the understanding of this lenition process, we present electromagnetic articulography and acoustic data from four speakers concernin...
null
null
johnson21_interspeech
Leveraging the Uniformity Framework to Examine Crosslinguistic Similarity for Long-Lag Stops in Spontaneous Cantonese-English Bilingual Speech
[ "Khia A. Johnson" ]
https://www.isca-archive.org/interspeech_2021/johnson21_interspeech.html
https://www.isca-archive.org/interspeech_2021/johnson21_interspeech.pdf
10.21437/Interspeech.2021-1780
2671-2675
@inproceedings{johnson21_interspeech, title = {{Leveraging the Uniformity Framework to Examine Crosslinguistic Similarity for Long-Lag Stops in Spontaneous Cantonese-English Bilingual Speech}}, author = {Khia A. Johnson}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2671--2675}, ...
While crosslinguistic influence is widespread in bilingual speech production, it is less clear which aspects of representation are shared across languages, if any. Most prior work examines phonetically distinct yet phonologically similar sounds, for which phonetic convergence suggests a cross-language link within indiv...
null
null
sivaraman21_interspeech
Personalized Speech Enhancement Through Self-Supervised Data Augmentation and Purification
[ "Aswin Sivaraman", "Sunwoo Kim", "Minje Kim" ]
https://www.isca-archive.org/interspeech_2021/sivaraman21_interspeech.html
https://www.isca-archive.org/interspeech_2021/sivaraman21_interspeech.pdf
10.21437/Interspeech.2021-1868
2676-2680
@inproceedings{sivaraman21_interspeech, title = {{Personalized Speech Enhancement Through Self-Supervised Data Augmentation and Purification}}, author = {Aswin Sivaraman and Sunwoo Kim and Minje Kim}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2676--2680}, doi = {10.2143...
Training personalized speech enhancement models is innately a no-shot learning problem due to privacy constraints and limited access to noise-free speech from the target user. If there is an abundance of unlabeled noisy speech from the test-time user, one may train a personalized speech enhancement model using self-sup...
2104.02018
title_snapshot
saddler21_interspeech
Speech Denoising with Auditory Models
[ "Mark R. Saddler", "Andrew Francl", "Jenelle Feather", "Kaizhi Qian", "Yang Zhang", "Josh H. McDermott" ]
https://www.isca-archive.org/interspeech_2021/saddler21_interspeech.html
https://www.isca-archive.org/interspeech_2021/saddler21_interspeech.pdf
10.21437/Interspeech.2021-1973
2681-2685
@inproceedings{saddler21_interspeech, title = {{Speech Denoising with Auditory Models}}, author = {Mark R. Saddler and Andrew Francl and Jenelle Feather and Kaizhi Qian and Yang Zhang and Josh H. McDermott}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2681--2685}, doi = {...
Contemporary speech enhancement predominantly relies on audio transforms that are trained to reconstruct a clean speech waveform. The development of high-performing neural network sound recognition systems has raised the possibility of using deep feature representations as ‘perceptual’ losses with which to train denois...
2011.10706
title_snapshot
eskimez21b_interspeech
Human Listening and Live Captioning: Multi-Task Training for Speech Enhancement
[ "Sefik Emre Eskimez", "Xiaofei Wang", "Min Tang", "Hemin Yang", "Zirun Zhu", "Zhuo Chen", "Huaming Wang", "Takuya Yoshioka" ]
https://www.isca-archive.org/interspeech_2021/eskimez21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/eskimez21b_interspeech.pdf
10.21437/Interspeech.2021-220
2686-2690
@inproceedings{eskimez21b_interspeech, title = {{Human Listening and Live Captioning: Multi-Task Training for Speech Enhancement}}, author = {Sefik Emre Eskimez and Xiaofei Wang and Min Tang and Hemin Yang and Zirun Zhu and Zhuo Chen and Huaming Wang and Takuya Yoshioka}, year = {2021}, booktitle = ...
With the surge of online meetings, it has become more critical than ever to provide high-quality speech audio and live captioning under various noise conditions. However, most monaural speech enhancement (SE) models introduce processing artifacts and thus degrade the performance of downstream tasks, including automatic...
2106.02896
title_snapshot
xu21g_interspeech
Multi-Stage Progressive Speech Enhancement Network
[ "Xinmeng Xu", "Yang Wang", "Dongxiang Xu", "Yiyuan Peng", "Cong Zhang", "Jie Jia", "Binbin Chen" ]
https://www.isca-archive.org/interspeech_2021/xu21g_interspeech.html
https://www.isca-archive.org/interspeech_2021/xu21g_interspeech.pdf
10.21437/Interspeech.2021-520
2691-2695
@inproceedings{xu21g_interspeech, title = {{Multi-Stage Progressive Speech Enhancement Network}}, author = {Xinmeng Xu and Yang Wang and Dongxiang Xu and Yiyuan Peng and Cong Zhang and Jie Jia and Binbin Chen}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2691--2695}, doi ...
Speech enhancement is a fundamental way to separate and generate clean speech from adverse environment where the received speech is seriously corrupted by noise. This paper applies a novel progressive network for speech enhancement by using multi-stage structure, where each stage contains a channel attention block foll...
null
null
chang21c_interspeech
Single-Channel Speech Enhancement Using Learnable Loss Mixup
[ "Oscar Chang", "Dung N. Tran", "Kazuhito Koishida" ]
https://www.isca-archive.org/interspeech_2021/chang21c_interspeech.html
https://www.isca-archive.org/interspeech_2021/chang21c_interspeech.pdf
10.21437/Interspeech.2021-859
2696-2700
@inproceedings{chang21c_interspeech, title = {{Single-Channel Speech Enhancement Using Learnable Loss Mixup}}, author = {Oscar Chang and Dung N. Tran and Kazuhito Koishida}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2696--2700}, doi = {10.21437/Interspeech.2021-859}, ...
Generalization remains a major problem in supervised learning of single-channel speech enhancement. In this work, we propose learnable loss mixup (LLM) , a simple and effortless training diagram, to improve the generalization of deep learning-based speech enhancement models. Loss mixup , of which learnable loss mixup i...
2312.17255
title_snapshot
zhang21q_interspeech
A Maximum Likelihood Approach to SNR-Progressive Learning Using Generalized Gaussian Distribution for LSTM-Based Speech Enhancement
[ "Xiao-Qi Zhang", "Jun Du", "Li Chai", "Chin-Hui Lee" ]
https://www.isca-archive.org/interspeech_2021/zhang21q_interspeech.html
https://www.isca-archive.org/interspeech_2021/zhang21q_interspeech.pdf
10.21437/Interspeech.2021-922
2701-2705
@inproceedings{zhang21q_interspeech, title = {{A Maximum Likelihood Approach to SNR-Progressive Learning Using Generalized Gaussian Distribution for LSTM-Based Speech Enhancement}}, author = {Xiao-Qi Zhang and Jun Du and Li Chai and Chin-Hui Lee}, year = {2021}, booktitle = {{Interspeech 2021}}, p...
A maximum likelihood (ML) approach to characterizing regression errors in each target layer of SNR progressive learning (PL) using long short-term memory (LSTM) networks is proposed to improve performances of speech enhancement at low SNR levels. Each LSTM layer is guided to learn an intermediate target with a specific...
null
null
agrawal21_interspeech
Whisper Speech Enhancement Using Joint Variational Autoencoder for Improved Speech Recognition
[ "Vikas Agrawal", "Shashi Kumar", "Shakti P. Rath" ]
https://www.isca-archive.org/interspeech_2021/agrawal21_interspeech.html
https://www.isca-archive.org/interspeech_2021/agrawal21_interspeech.pdf
10.21437/Interspeech.2021-953
2706-2710
@inproceedings{agrawal21_interspeech, title = {{Whisper Speech Enhancement Using Joint Variational Autoencoder for Improved Speech Recognition}}, author = {Vikas Agrawal and Shashi Kumar and Shakti P. Rath}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2706--2710}, doi = {...
Whispering is the natural choice of communication when one wants to interact quietly and privately. Due to vast differences in acoustic characteristics of whisper and natural speech, there is drastic degradation in the performance of whisper speech when decoded by the Automatic Speech Recognition (ASR) system trained o...
null
null
lee21d_interspeech
DEMUCS-Mobile : On-Device Lightweight Speech Enhancement
[ "Lukas Lee", "Youna Ji", "Minjae Lee", "Min-Seok Choi" ]
https://www.isca-archive.org/interspeech_2021/lee21d_interspeech.html
https://www.isca-archive.org/interspeech_2021/lee21d_interspeech.pdf
10.21437/Interspeech.2021-1025
2711-2715
@inproceedings{lee21d_interspeech, title = {{DEMUCS-Mobile : On-Device Lightweight Speech Enhancement}}, author = {Lukas Lee and Youna Ji and Minjae Lee and Min-Seok Choi}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2711--2715}, doi = {10.21437/Interspeech.2021-1025}, ...
As the importance of speech enhancement for real-world application increases, the compactness of the model is also becoming a crucial study. In this paper, we present compression techniques to reduce the model size and applied them to the state-of-the-art real-time speech enhancement system. We successfully reduce the ...
null
null
kashyap21_interspeech
Speech Denoising Without Clean Training Data: A Noise2Noise Approach
[ "Madhav Mahesh Kashyap", "Anuj Tambwekar", "Krishnamoorthy Manohara", "S. Natarajan" ]
https://www.isca-archive.org/interspeech_2021/kashyap21_interspeech.html
https://www.isca-archive.org/interspeech_2021/kashyap21_interspeech.pdf
10.21437/Interspeech.2021-1130
2716-2720
@inproceedings{kashyap21_interspeech, title = {{Speech Denoising Without Clean Training Data: A Noise2Noise Approach}}, author = {Madhav Mahesh Kashyap and Anuj Tambwekar and Krishnamoorthy Manohara and S. Natarajan}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2716--2720}, doi...
This paper tackles the problem of the heavy dependence of clean speech data required by deep learning based audio-denoising methods by showing that it is possible to train deep speech denoising networks using only noisy speech samples. Conventional wisdom dictates that in order to achieve good speech denoising performa...
2104.03838
title_snapshot
dang21_interspeech
Improved Speech Enhancement Using a Complex-Domain GAN with Fused Time-Domain and Time-Frequency Domain Constraints
[ "Feng Dang", "Pengyuan Zhang", "Hangting Chen" ]
https://www.isca-archive.org/interspeech_2021/dang21_interspeech.html
https://www.isca-archive.org/interspeech_2021/dang21_interspeech.pdf
10.21437/Interspeech.2021-1134
2721-2725
@inproceedings{dang21_interspeech, title = {{Improved Speech Enhancement Using a Complex-Domain GAN with Fused Time-Domain and Time-Frequency Domain Constraints}}, author = {Feng Dang and Pengyuan Zhang and Hangting Chen}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2721--2725}, ...
Complex-domain models have achieved promising results for speech enhancement (SE) tasks. Some complex-domain models consider only time-frequency (T-F) domain constraints and do not take advantage of the information at the time-domain waveform level. Some complex-domain models consider only time-domain constraints and d...
null
null
zhang21r_interspeech
Speech Enhancement with Topology-Enhanced Generative Adversarial Networks (GANs)
[ "Xudong Zhang", "Liang Zhao", "Feng Gu" ]
https://www.isca-archive.org/interspeech_2021/zhang21r_interspeech.html
https://www.isca-archive.org/interspeech_2021/zhang21r_interspeech.pdf
10.21437/Interspeech.2021-1411
2726-2730
@inproceedings{zhang21r_interspeech, title = {{Speech Enhancement with Topology-Enhanced Generative Adversarial Networks (GANs)}}, author = {Xudong Zhang and Liang Zhao and Feng Gu}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2726--2730}, doi = {10.21437/Interspeech.2021...
Speech enhancement is one of the effective approaches in improving speech quality. Neural network models have been widely used in speech enhancement, such as recurrent neural networks (RNNs), long short-term memory networks (LSTMs), and generative adversarial networks (GANs). However, some of them either handle the spe...
null
null
bu21_interspeech
Learning Speech Structure to Improve Time-Frequency Masks
[ "Suliang Bu", "Yunxin Zhao", "Shaojun Wang", "Mei Han" ]
https://www.isca-archive.org/interspeech_2021/bu21_interspeech.html
https://www.isca-archive.org/interspeech_2021/bu21_interspeech.pdf
10.21437/Interspeech.2021-1859
2731-2735
@inproceedings{bu21_interspeech, title = {{Learning Speech Structure to Improve Time-Frequency Masks}}, author = {Suliang Bu and Yunxin Zhao and Shaojun Wang and Mei Han}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2731--2735}, doi = {10.21437/Interspeech.2021-1859}, i...
Time-frequency (TF) masks are widely used in speech enhancement (SE). However, accurately estimating TF masks from noisy speech remains a challenge to both statistical or neural network approaches. Statistical model-based mask estimation usually depends on a good parameter initialization, while NN-based mask estimation...
null
null
kim21h_interspeech
SE-Conformer: Time-Domain Speech Enhancement Using Conformer
[ "Eesung Kim", "Hyeji Seo" ]
https://www.isca-archive.org/interspeech_2021/kim21h_interspeech.html
https://www.isca-archive.org/interspeech_2021/kim21h_interspeech.pdf
10.21437/Interspeech.2021-2207
2736-2740
@inproceedings{kim21h_interspeech, title = {{SE-Conformer: Time-Domain Speech Enhancement Using Conformer}}, author = {Eesung Kim and Hyeji Seo}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2736--2740}, doi = {10.21437/Interspeech.2021-2207}, issn = {2958-1796}, }
Convolution-augmented transformer (conformer) has recently shown competitive results in speech-domain applications, such as automatic speech recognition, continuous speech separation, and sound event detection. Conformer can capture both the short and long-term temporal sequence information by attending to the whole se...
null
null
kongthaworn21_interspeech
Spectral and Latent Speech Representation Distortion for TTS Evaluation
[ "Thananchai Kongthaworn", "Burin Naowarat", "Ekapol Chuangsuwanich" ]
https://www.isca-archive.org/interspeech_2021/kongthaworn21_interspeech.html
https://www.isca-archive.org/interspeech_2021/kongthaworn21_interspeech.pdf
10.21437/Interspeech.2021-2258
2741-2745
@inproceedings{kongthaworn21_interspeech, title = {{Spectral and Latent Speech Representation Distortion for TTS Evaluation}}, author = {Thananchai Kongthaworn and Burin Naowarat and Ekapol Chuangsuwanich}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2741--2745}, doi = {1...
One of the main problems in the development of text-to-speech (TTS) systems is its reliance on subjective measures, typically the Mean Opinion Score (MOS). MOS requires a large number of people to reliably rate each utterance, making the development process slow and expensive. Recent research on speech quality assessme...
null
null
valentinibotinhao21_interspeech
Detection and Analysis of Attention Errors in Sequence-to-Sequence Text-to-Speech
[ "Cassia Valentini-Botinhao", "Simon King" ]
https://www.isca-archive.org/interspeech_2021/valentinibotinhao21_interspeech.html
https://www.isca-archive.org/interspeech_2021/valentinibotinhao21_interspeech.pdf
10.21437/Interspeech.2021-286
2746-2750
@inproceedings{valentinibotinhao21_interspeech, title = {{Detection and Analysis of Attention Errors in Sequence-to-Sequence Text-to-Speech}}, author = {Cassia Valentini-Botinhao and Simon King}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2746--2750}, doi = {10.21437/Int...
Sequence-to-sequence speech synthesis models are notorious for gross errors such as skipping and repetition, commonly associated with failures in the attention mechanism. While a lot has been done to improve attention and decrease errors, this paper focuses instead on automatic error detection and analysis. We evaluate...
null
null
zandie21_interspeech
RyanSpeech: A Corpus for Conversational Text-to-Speech Synthesis
[ "Rohola Zandie", "Mohammad H. Mahoor", "Julia Madsen", "Eshrat S. Emamian" ]
https://www.isca-archive.org/interspeech_2021/zandie21_interspeech.html
https://www.isca-archive.org/interspeech_2021/zandie21_interspeech.pdf
10.21437/Interspeech.2021-341
2751-2755
@inproceedings{zandie21_interspeech, title = {{RyanSpeech: A Corpus for Conversational Text-to-Speech Synthesis}}, author = {Rohola Zandie and Mohammad H. Mahoor and Julia Madsen and Eshrat S. Emamian}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2751--2755}, doi = {10.21...
This paper introduces RyanSpeech , a new speech corpus for research on automated text-to-speech (TTS) systems. Publicly available TTS corpora are often noisy, recorded with multiple speakers, or lack quality male speech data. In order to meet the need for a high quality, publicly available male speech corpus within the...
2106.08468
title_snapshot
shi21c_interspeech
AISHELL-3: A Multi-Speaker Mandarin TTS Corpus
[ "Yao Shi", "Hui Bu", "Xin Xu", "Shaoji Zhang", "Ming Li" ]
https://www.isca-archive.org/interspeech_2021/shi21c_interspeech.html
https://www.isca-archive.org/interspeech_2021/shi21c_interspeech.pdf
10.21437/Interspeech.2021-755
2756-2760
@inproceedings{shi21c_interspeech, title = {{AISHELL-3: A Multi-Speaker Mandarin TTS Corpus}}, author = {Yao Shi and Hui Bu and Xin Xu and Shaoji Zhang and Ming Li}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2756--2760}, doi = {10.21437/Interspeech.2021-755}, issn ...
In this paper, we present AISHELL-3, a large-scale multi-speaker Mandarin speech corpus which could be used to train multi-speaker Text-To-Speech (TTS) systems. The corpus contains roughly 85 hours of emotion-neutral recordings spanning across 218 native Chinese mandarin speakers. Their auxiliary attributes such as gen...
2010.11567
title_judge
eng21_interspeech
Comparing Speech Enhancement Techniques for Voice Adaptation-Based Speech Synthesis
[ "Nicholas Eng", "C.T. Justine Hui", "Yusuke Hioka", "Catherine I. Watson" ]
https://www.isca-archive.org/interspeech_2021/eng21_interspeech.html
https://www.isca-archive.org/interspeech_2021/eng21_interspeech.pdf
10.21437/Interspeech.2021-800
2761-2765
@inproceedings{eng21_interspeech, title = {{Comparing Speech Enhancement Techniques for Voice Adaptation-Based Speech Synthesis}}, author = {Nicholas Eng and C.T. Justine Hui and Yusuke Hioka and Catherine I. Watson}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2761--2765}, doi...
This study investigates the use of speech enhancement techniques in creating text-to-speech voices with degraded or noisy speech. A number of synthetic voices were created using speech that was first degraded by different noise types at various signal-to-noise ratios (SNRs), then enhanced through four speech enhancemen...
null
null
cui21c_interspeech
EMOVIE: A Mandarin Emotion Speech Dataset with a Simple Emotional Text-to-Speech Model
[ "Chenye Cui", "Yi Ren", "Jinglin Liu", "Feiyang Chen", "Rongjie Huang", "Ming Lei", "Zhou Zhao" ]
https://www.isca-archive.org/interspeech_2021/cui21c_interspeech.html
https://www.isca-archive.org/interspeech_2021/cui21c_interspeech.pdf
10.21437/Interspeech.2021-1148
2766-2770
@inproceedings{cui21c_interspeech, title = {{EMOVIE: A Mandarin Emotion Speech Dataset with a Simple Emotional Text-to-Speech Model}}, author = {Chenye Cui and Yi Ren and Jinglin Liu and Feiyang Chen and Rongjie Huang and Ming Lei and Zhou Zhao}, year = {2021}, booktitle = {{Interspeech 2021}}, pa...
Recently, there has been an increasing interest in neural speech synthesis. While the deep neural network achieves the state-of-the-art result in text-to-speech (TTS) tasks, how to generate a more emotional and more expressive speech is becoming a new challenge to researchers due to the scarcity of high-quality emotion...
2106.09317
title_snapshot
rallabandi21_interspeech
Perception of Social Speaker Characteristics in Synthetic Speech
[ "Sai Sirisha Rallabandi", "Abhinav Bharadwaj", "Babak Naderi", "Sebastian Möller" ]
https://www.isca-archive.org/interspeech_2021/rallabandi21_interspeech.html
https://www.isca-archive.org/interspeech_2021/rallabandi21_interspeech.pdf
10.21437/Interspeech.2021-1229
2771-2775
@inproceedings{rallabandi21_interspeech, title = {{Perception of Social Speaker Characteristics in Synthetic Speech}}, author = {Sai Sirisha Rallabandi and Abhinav Bharadwaj and Babak Naderi and Sebastian Möller}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2771--2775}, doi ...
With the improved computational abilities, the usage of chatbots and conversational agents has become more prevalent. Therefore, it is essential that these agents exhibit certain social speaker characteristics in the generated speech. In this paper, we study the perception of such speaker characteristics in two commerc...
null
null
bakhturina21_interspeech
Hi-Fi Multi-Speaker English TTS Dataset
[ "Evelina Bakhturina", "Vitaly Lavrukhin", "Boris Ginsburg", "Yang Zhang" ]
https://www.isca-archive.org/interspeech_2021/bakhturina21_interspeech.html
https://www.isca-archive.org/interspeech_2021/bakhturina21_interspeech.pdf
10.21437/Interspeech.2021-1599
2776-2780
@inproceedings{bakhturina21_interspeech, title = {{Hi-Fi Multi-Speaker English TTS Dataset}}, author = {Evelina Bakhturina and Vitaly Lavrukhin and Boris Ginsburg and Yang Zhang}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2776--2780}, doi = {10.21437/Interspeech.2021-15...
This paper introduces a new multi-speaker English dataset for training text-to-speech models. The dataset is based on LibriVox audiobooks and Project Gutenberg texts, both in the public domain. The new dataset contains about 292 hours of speech from 10 speakers with at least 17 hours per speaker sampled at 44.1 kHz. To...
2104.01497
title_snapshot
tseng21b_interspeech
Utilizing Self-Supervised Representations for MOS Prediction
[ "Wei-Cheng Tseng", "Chien-yu Huang", "Wei-Tsung Kao", "Yist Y. Lin", "Hung-yi Lee" ]
https://www.isca-archive.org/interspeech_2021/tseng21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/tseng21b_interspeech.pdf
10.21437/Interspeech.2021-2013
2781-2785
@inproceedings{tseng21b_interspeech, title = {{Utilizing Self-Supervised Representations for MOS Prediction}}, author = {Wei-Cheng Tseng and Chien-yu Huang and Wei-Tsung Kao and Yist Y. Lin and Hung-yi Lee}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2781--2785}, doi = {...
Speech quality assessment has been a critical issue in speech processing for decades. Existing automatic evaluations usually require clean references or parallel ground truth data, which is infeasible when the amount of data soars. Subjective tests, on the other hand, do not need any additional clean or parallel data a...
2104.03017
title_snapshot
mussakhojayeva21_interspeech
KazakhTTS: An Open-Source Kazakh Text-to-Speech Synthesis Dataset
[ "Saida Mussakhojayeva", "Aigerim Janaliyeva", "Almas Mirzakhmetov", "Yerbolat Khassanov", "Huseyin Atakan Varol" ]
https://www.isca-archive.org/interspeech_2021/mussakhojayeva21_interspeech.html
https://www.isca-archive.org/interspeech_2021/mussakhojayeva21_interspeech.pdf
10.21437/Interspeech.2021-2124
2786-2790
@inproceedings{mussakhojayeva21_interspeech, title = {{KazakhTTS: An Open-Source Kazakh Text-to-Speech Synthesis Dataset}}, author = {Saida Mussakhojayeva and Aigerim Janaliyeva and Almas Mirzakhmetov and Yerbolat Khassanov and Huseyin Atakan Varol}, year = {2021}, booktitle = {{Interspeech 2021}}, ...
This paper introduces a high-quality open-source speech synthesis dataset for Kazakh, a low-resource language spoken by over 13 million people worldwide. The dataset consists of about 93 hours of transcribed audio recordings spoken by two professional speakers (female and male). It is the first publicly available large...
2104.08459
title_snapshot
taylor21_interspeech
Confidence Intervals for ASR-Based TTS Evaluation
[ "Jason Taylor", "Korin Richmond" ]
https://www.isca-archive.org/interspeech_2021/taylor21_interspeech.html
https://www.isca-archive.org/interspeech_2021/taylor21_interspeech.pdf
10.21437/Interspeech.2021-2203
2791-2795
@inproceedings{taylor21_interspeech, title = {{Confidence Intervals for ASR-Based TTS Evaluation}}, author = {Jason Taylor and Korin Richmond}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2791--2795}, doi = {10.21437/Interspeech.2021-2203}, issn = {2958-1796}, }
Automatic speech recognition (ASR) is increasingly used to evaluate the intelligibility of text-to-speech synthesis (TTS). ASR is less costly than traditional listening tests, but questions remain about its reliability. We re-evaluate the Blizzard Challenge’s intelligibility tasks in English since 2011 using ASR. Re-an...
null
null
reddy21_interspeech
INTERSPEECH 2021 Deep Noise Suppression Challenge
[ "Chandan K.A. Reddy", "Harishchandra Dubey", "Kazuhito Koishida", "Arun Nair", "Vishak Gopal", "Ross Cutler", "Sebastian Braun", "Hannes Gamper", "Robert Aichner", "Sriram Srinivasan" ]
https://www.isca-archive.org/interspeech_2021/reddy21_interspeech.html
https://www.isca-archive.org/interspeech_2021/reddy21_interspeech.pdf
10.21437/Interspeech.2021-1609
2796-2800
@inproceedings{reddy21_interspeech, title = {{INTERSPEECH 2021 Deep Noise Suppression Challenge}}, author = {Chandan K.A. Reddy and Harishchandra Dubey and Kazuhito Koishida and Arun Nair and Vishak Gopal and Ross Cutler and Sebastian Braun and Hannes Gamper and Robert Aichner and Sriram Srinivasan}, year ...
The Deep Noise Suppression (DNS) challenge was designed to unify the research efforts in the area of noise suppression targeted for human perception. We recently organized a DNS challenge special session at INTERSPEECH 2020 and ICASSP 2021. We open-sourced training and test datasets for the wideband scenario along with...
2101.01902
title_snapshot
li21g_interspeech
A Simultaneous Denoising and Dereverberation Framework with Target Decoupling
[ "Andong Li", "Wenzhe Liu", "Xiaoxue Luo", "Guochen Yu", "Chengshi Zheng", "Xiaodong Li" ]
https://www.isca-archive.org/interspeech_2021/li21g_interspeech.html
https://www.isca-archive.org/interspeech_2021/li21g_interspeech.pdf
10.21437/Interspeech.2021-1137
2801-2805
@inproceedings{li21g_interspeech, title = {{A Simultaneous Denoising and Dereverberation Framework with Target Decoupling}}, author = {Andong Li and Wenzhe Liu and Xiaoxue Luo and Guochen Yu and Chengshi Zheng and Xiaodong Li}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2801--28...
Background noise and room reverberation are regarded as two major factors to degrade the subjective speech quality. In this paper, we propose an integrated framework to address simultaneous denoising and dereverberation under complicated scenario environments. It adopts a chain optimization strategy and designs four su...
2106.12743
title_snapshot
xu21h_interspeech
Deep Noise Suppression with Non-Intrusive PESQNet Supervision Enabling the Use of Real Training Data
[ "Ziyi Xu", "Maximilian Strake", "Tim Fingscheidt" ]
https://www.isca-archive.org/interspeech_2021/xu21h_interspeech.html
https://www.isca-archive.org/interspeech_2021/xu21h_interspeech.pdf
10.21437/Interspeech.2021-936
2806-2810
@inproceedings{xu21h_interspeech, title = {{Deep Noise Suppression with Non-Intrusive PESQNet Supervision Enabling the Use of Real Training Data}}, author = {Ziyi Xu and Maximilian Strake and Tim Fingscheidt}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2806--2810}, doi =...
Data-driven speech enhancement employing deep neural networks (DNNs) can provide state-of-the-art performance even in the presence of non-stationary noise. During the training process, most of the speech enhancement neural networks are trained in a fully supervised way with losses requiring noisy speech to be synthesiz...
2103.17088
title_snapshot
le21b_interspeech
DPCRN: Dual-Path Convolution Recurrent Network for Single Channel Speech Enhancement
[ "Xiaohuai Le", "Hongsheng Chen", "Kai Chen", "Jing Lu" ]
https://www.isca-archive.org/interspeech_2021/le21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/le21b_interspeech.pdf
10.21437/Interspeech.2021-296
2811-2815
@inproceedings{le21b_interspeech, title = {{DPCRN: Dual-Path Convolution Recurrent Network for Single Channel Speech Enhancement}}, author = {Xiaohuai Le and Hongsheng Chen and Kai Chen and Jing Lu}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2811--2815}, doi = {10.21437...
The dual-path RNN (DPRNN) was proposed to more effectively model extremely long sequences for speech separation in the time domain. By splitting long sequences to smaller chunks and applying intra-chunk and inter-chunk RNNs, the DPRNN reached promising performance in speech separation with a limited model size. In this...
2107.05429
title_snapshot
lv21_interspeech
DCCRN+: Channel-Wise Subband DCCRN with SNR Estimation for Speech Enhancement
[ "Shubo Lv", "Yanxin Hu", "Shimin Zhang", "Lei Xie" ]
https://www.isca-archive.org/interspeech_2021/lv21_interspeech.html
https://www.isca-archive.org/interspeech_2021/lv21_interspeech.pdf
10.21437/Interspeech.2021-1482
2816-2820
@inproceedings{lv21_interspeech, title = {{DCCRN+: Channel-Wise Subband DCCRN with SNR Estimation for Speech Enhancement}}, author = {Shubo Lv and Yanxin Hu and Shimin Zhang and Lei Xie}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2816--2820}, doi = {10.21437/Interspeech...
Deep complex convolution recurrent network (DCCRN), which extends CRN with complex structure, has achieved superior performance in MOS evaluation in Interspeech 2020 deep noise suppression challenge (DNS2020). This paper further extends DCCRN with the following significant revisions. We first extend the model to sub-ba...
2106.08672
title_snapshot
zhang21s_interspeech
DBNet: A Dual-Branch Network Architecture Processing on Spectrum and Waveform for Single-Channel Speech Enhancement
[ "Kanghao Zhang", "Shulin He", "Hao Li", "Xueliang Zhang" ]
https://www.isca-archive.org/interspeech_2021/zhang21s_interspeech.html
https://www.isca-archive.org/interspeech_2021/zhang21s_interspeech.pdf
10.21437/Interspeech.2021-1042
2821-2825
@inproceedings{zhang21s_interspeech, title = {{DBNet: A Dual-Branch Network Architecture Processing on Spectrum and Waveform for Single-Channel Speech Enhancement}}, author = {Kanghao Zhang and Shulin He and Hao Li and Xueliang Zhang}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {...
In real acoustic environment, speech enhancement is an arduous task to improve the quality and intelligibility of speech interfered by background noise and reverberation. Over the past years, deep learning has shown great potential on speech enhancement. In this paper, we propose a novel real-time framework called DBNe...
2105.02436
title_snapshot
zhang21t_interspeech
Low-Delay Speech Enhancement Using Perceptually Motivated Target and Loss
[ "Xu Zhang", "Xinlei Ren", "Xiguang Zheng", "Lianwu Chen", "Chen Zhang", "Liang Guo", "Bing Yu" ]
https://www.isca-archive.org/interspeech_2021/zhang21t_interspeech.html
https://www.isca-archive.org/interspeech_2021/zhang21t_interspeech.pdf
10.21437/Interspeech.2021-1410
2826-2830
@inproceedings{zhang21t_interspeech, title = {{Low-Delay Speech Enhancement Using Perceptually Motivated Target and Loss}}, author = {Xu Zhang and Xinlei Ren and Xiguang Zheng and Lianwu Chen and Chen Zhang and Liang Guo and Bing Yu}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2...
Speech enhancement approaches based on deep neural network have outperformed the traditional signal processing methods. This paper presents a low-delay speech enhancement method that employs a new perceptually motivated training target and loss function. The proposed approach can achieve similar speech enhancement perf...
null
null
oostermeijer21_interspeech
Lightweight Causal Transformer with Local Self-Attention for Real-Time Speech Enhancement
[ "Koen Oostermeijer", "Qing Wang", "Jun Du" ]
https://www.isca-archive.org/interspeech_2021/oostermeijer21_interspeech.html
https://www.isca-archive.org/interspeech_2021/oostermeijer21_interspeech.pdf
10.21437/Interspeech.2021-668
2831-2835
@inproceedings{oostermeijer21_interspeech, title = {{Lightweight Causal Transformer with Local Self-Attention for Real-Time Speech Enhancement}}, author = {Koen Oostermeijer and Qing Wang and Jun Du}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2831--2835}, doi = {10.2143...
In this paper, we describe a novel speech enhancement transformer architecture. The model uses local causal self-attention, which makes it lightweight and therefore particularly well-suited for real-time speech enhancement in computation resource-limited environments. In addition, we provide several ablation studies th...
null
null
ristea21_interspeech
Self-Paced Ensemble Learning for Speech and Audio Classification
[ "Nicolae-Cătălin Ristea", "Radu Tudor Ionescu" ]
https://www.isca-archive.org/interspeech_2021/ristea21_interspeech.html
https://www.isca-archive.org/interspeech_2021/ristea21_interspeech.pdf
10.21437/Interspeech.2021-155
2836-2840
@inproceedings{ristea21_interspeech, title = {{Self-Paced Ensemble Learning for Speech and Audio Classification}}, author = {Nicolae-Cătălin Ristea and Radu Tudor Ionescu}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2836--2840}, doi = {10.21437/Interspeech.2021-155}, i...
Combining multiple machine learning models into an ensemble is known to provide superior performance levels compared to the individual components forming the ensemble. This is because models can complement each other in taking better decisions. Instead of just combining the models, we propose a self-paced ensemble lear...
2103.11988
title_snapshot
kojima21_interspeech
Knowledge Distillation for Streaming Transformer–Transducer
[ "Atsushi Kojima" ]
https://www.isca-archive.org/interspeech_2021/kojima21_interspeech.html
https://www.isca-archive.org/interspeech_2021/kojima21_interspeech.pdf
10.21437/Interspeech.2021-175
2841-2845
@inproceedings{kojima21_interspeech, title = {{Knowledge Distillation for Streaming Transformer–Transducer}}, author = {Atsushi Kojima}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2841--2845}, doi = {10.21437/Interspeech.2021-175}, issn = {2958-1796}, }
We explore knowledge distillation methods from nonstreaming to streaming Transformer–Transducer (T–T) models. Streaming T–T truncates future context. It leads to recognition quality degradation compared with the original T–T. In this work, we explore knowledge distillation, which minimizes internal representations in a...
null
null
lohrenz21_interspeech
Multi-Encoder Learning and Stream Fusion for Transformer-Based End-to-End Automatic Speech Recognition
[ "Timo Lohrenz", "Zhengyang Li", "Tim Fingscheidt" ]
https://www.isca-archive.org/interspeech_2021/lohrenz21_interspeech.html
https://www.isca-archive.org/interspeech_2021/lohrenz21_interspeech.pdf
10.21437/Interspeech.2021-555
2846-2850
@inproceedings{lohrenz21_interspeech, title = {{Multi-Encoder Learning and Stream Fusion for Transformer-Based End-to-End Automatic Speech Recognition}}, author = {Timo Lohrenz and Zhengyang Li and Tim Fingscheidt}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2846--2850}, doi ...
Stream fusion, also known as system combination, is a common technique in automatic speech recognition for traditional hybrid hidden Markov model approaches, yet mostly unexplored for modern deep neural network end-to-end model architectures. Here, we investigate various fusion techniques for the all-attention-based en...
2104.00120
title_snapshot
zaiem21_interspeech
Conditional Independence for Pretext Task Selection in Self-Supervised Speech Representation Learning
[ "Salah Zaiem", "Titouan Parcollet", "Slim Essid" ]
https://www.isca-archive.org/interspeech_2021/zaiem21_interspeech.html
https://www.isca-archive.org/interspeech_2021/zaiem21_interspeech.pdf
10.21437/Interspeech.2021-1027
2851-2855
@inproceedings{zaiem21_interspeech, title = {{Conditional Independence for Pretext Task Selection in Self-Supervised Speech Representation Learning}}, author = {Salah Zaiem and Titouan Parcollet and Slim Essid}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2851--2855}, doi ...
Through solving pretext tasks, self-supervised learning (SSL) leverages unlabeled data to extract useful latent representations replacing traditional input features in the downstream task. A common pretext task consists in pretraining a SSL model on pseudo-labels derived from the original signal. This technique is part...
2104.07388
title_snapshot
zeineldeen21_interspeech
Investigating Methods to Improve Language Model Integration for Attention-Based Encoder-Decoder ASR Models
[ "Mohammad Zeineldeen", "Aleksandr Glushko", "Wilfried Michel", "Albert Zeyer", "Ralf Schlüter", "Hermann Ney" ]
https://www.isca-archive.org/interspeech_2021/zeineldeen21_interspeech.html
https://www.isca-archive.org/interspeech_2021/zeineldeen21_interspeech.pdf
10.21437/Interspeech.2021-1255
2856-2860
@inproceedings{zeineldeen21_interspeech, title = {{Investigating Methods to Improve Language Model Integration for Attention-Based Encoder-Decoder ASR Models}}, author = {Mohammad Zeineldeen and Aleksandr Glushko and Wilfried Michel and Albert Zeyer and Ralf Schlüter and Hermann Ney}, year = {2021}, ...
Attention-based encoder-decoder (AED) models learn an implicit internal language model (ILM) from the training transcriptions. The integration with an external LM trained on much more unpaired text usually leads to better performance. A Bayesian interpretation as in the hybrid autoregressive transducer (HAT) suggests d...
2104.05544
title_snapshot
vyas21b_interspeech
Comparing CTC and LFMMI for Out-of-Domain Adaptation of wav2vec 2.0 Acoustic Model
[ "Apoorv Vyas", "Srikanth Madikeri", "Hervé Bourlard" ]
https://www.isca-archive.org/interspeech_2021/vyas21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/vyas21b_interspeech.pdf
10.21437/Interspeech.2021-1683
2861-2865
@inproceedings{vyas21b_interspeech, title = {{Comparing CTC and LFMMI for Out-of-Domain Adaptation of wav2vec 2.0 Acoustic Model}}, author = {Apoorv Vyas and Srikanth Madikeri and Hervé Bourlard}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2861--2865}, doi = {10.21437/In...
In this work, we investigate if the wav2vec 2.0 self-supervised pretraining helps mitigate the overfitting issues with connectionist temporal classification (CTC) training to reduce its performance gap with flat-start lattice-free MMI (E2E-LFMMI) for automatic speech recognition with limited training data. Towards that...
2104.02558
title_snapshot
moine21_interspeech
Speaker Attentive Speech Emotion Recognition
[ "Clément Le Moine", "Nicolas Obin", "Axel Roebel" ]
https://www.isca-archive.org/interspeech_2021/moine21_interspeech.html
https://www.isca-archive.org/interspeech_2021/moine21_interspeech.pdf
10.21437/Interspeech.2021-573
2866-2870
@inproceedings{moine21_interspeech, title = {{Speaker Attentive Speech Emotion Recognition}}, author = {Clément Le Moine and Nicolas Obin and Axel Roebel}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2866--2870}, doi = {10.21437/Interspeech.2021-573}, issn = {2958-...
Speech Emotion Recognition (SER) task has known significant improvements over the last years with the advent of Deep Neural Networks (DNNs). However, even the most successful methods are still rather failing when adaptation to specific speakers and scenarios is needed, inevitably leading to poorer performances when com...
2104.07288
title_snapshot
leem21_interspeech
Separation of Emotional and Reconstruction Embeddings on Ladder Network to Improve Speech Emotion Recognition Robustness in Noisy Conditions
[ "Seong-Gyun Leem", "Daniel Fulford", "Jukka-Pekka Onnela", "David Gard", "Carlos Busso" ]
https://www.isca-archive.org/interspeech_2021/leem21_interspeech.html
https://www.isca-archive.org/interspeech_2021/leem21_interspeech.pdf
10.21437/Interspeech.2021-1438
2871-2875
@inproceedings{leem21_interspeech, title = {{Separation of Emotional and Reconstruction Embeddings on Ladder Network to Improve Speech Emotion Recognition Robustness in Noisy Conditions}}, author = {Seong-Gyun Leem and Daniel Fulford and Jukka-Pekka Onnela and David Gard and Carlos Busso}, year = {202...
When speech emotion recognition (SER) is applied in an actual application, the system should be able to cope with audio acquired in a noisy, unconstrained environment. Most studies on noise-robust SER require a parallel dataset with emotion labels, which is impractical to collect, or use speech with artificially added ...
null
null
georgiou21_interspeech
M: MultiModal Masking Applied to Sentiment Analysis
[ "Efthymios Georgiou", "Georgios Paraskevopoulos", "Alexandros Potamianos" ]
https://www.isca-archive.org/interspeech_2021/georgiou21_interspeech.html
https://www.isca-archive.org/interspeech_2021/georgiou21_interspeech.pdf
10.21437/Interspeech.2021-1739
2876-2880
@inproceedings{georgiou21_interspeech, title = {{M3: MultiModal Masking Applied to Sentiment Analysis}}, author = {Efthymios Georgiou and Georgios Paraskevopoulos and Alexandros Potamianos}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2876--2880}, doi = {10.21437/Interspe...
A common issue when training multimodal architectures is that not all modalities contribute equally to the model’s prediction and the network tends to over-rely on the strongest modality. In this work, we present M 3 , a training procedure based on modality masking for deep multimodal architectures. During network trai...
null
null
klejch21_interspeech
The CSTR System for Multilingual and Code-Switching ASR Challenges for Low Resource Indian Languages
[ "Ondřej Klejch", "Electra Wallington", "Peter Bell" ]
https://www.isca-archive.org/interspeech_2021/klejch21_interspeech.html
https://www.isca-archive.org/interspeech_2021/klejch21_interspeech.pdf
10.21437/Interspeech.2021-1035
2881-2885
@inproceedings{klejch21_interspeech, title = {{The CSTR System for Multilingual and Code-Switching ASR Challenges for Low Resource Indian Languages}}, author = {Ondřej Klejch and Electra Wallington and Peter Bell}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2881--2885}, doi ...
This paper describes the CSTR submission to the Multilingual and Code-Switching ASR Challenges at Interspeech 2021. For the multilingual track of the challenge, we trained a multilingual CNN-TDNN acoustic model for Gujarati, Hindi, Marathi, Odia, Tamil and Telugu and subsequently fine-tuned the model on monolingual tra...
null
null
zhou21d_interspeech
Acoustic Data-Driven Subword Modeling for End-to-End Speech Recognition
[ "Wei Zhou", "Mohammad Zeineldeen", "Zuoyun Zheng", "Ralf Schlüter", "Hermann Ney" ]
https://www.isca-archive.org/interspeech_2021/zhou21d_interspeech.html
https://www.isca-archive.org/interspeech_2021/zhou21d_interspeech.pdf
10.21437/Interspeech.2021-1623
2886-2890
@inproceedings{zhou21d_interspeech, title = {{Acoustic Data-Driven Subword Modeling for End-to-End Speech Recognition}}, author = {Wei Zhou and Mohammad Zeineldeen and Zuoyun Zheng and Ralf Schlüter and Hermann Ney}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2886--2890}, doi ...
Subword units are commonly used for end-to-end automatic speech recognition (ASR), while a fully acoustic-oriented subword modeling approach is somewhat missing. We propose an acoustic data-driven subword modeling (ADSM) approach that adapts the advantages of several text-based and acoustic-based subword methods into o...
2104.09106
title_snapshot
zhou21e_interspeech
Equivalence of Segmental and Neural Transducer Modeling: A Proof of Concept
[ "Wei Zhou", "Albert Zeyer", "André Merboldt", "Ralf Schlüter", "Hermann Ney" ]
https://www.isca-archive.org/interspeech_2021/zhou21e_interspeech.html
https://www.isca-archive.org/interspeech_2021/zhou21e_interspeech.pdf
10.21437/Interspeech.2021-1671
2891-2895
@inproceedings{zhou21e_interspeech, title = {{Equivalence of Segmental and Neural Transducer Modeling: A Proof of Concept}}, author = {Wei Zhou and Albert Zeyer and André Merboldt and Ralf Schlüter and Hermann Ney}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2891--2895}, doi ...
With the advent of direct models in automatic speech recognition (ASR), the formerly prevalent frame-wise acoustic modeling based on hidden Markov models (HMM) diversified into a number of modeling architectures like encoder-decoder attention models, transducer models and segmental models (direct HMM). While transducer...
2104.06104
title_snapshot
khosravani21_interspeech
Modeling Dialectal Variation for Swiss German Automatic Speech Recognition
[ "Abbas Khosravani", "Philip N. Garner", "Alexandros Lazaridis" ]
https://www.isca-archive.org/interspeech_2021/khosravani21_interspeech.html
https://www.isca-archive.org/interspeech_2021/khosravani21_interspeech.pdf
10.21437/Interspeech.2021-1735
2896-2900
@inproceedings{khosravani21_interspeech, title = {{Modeling Dialectal Variation for Swiss German Automatic Speech Recognition}}, author = {Abbas Khosravani and Philip N. Garner and Alexandros Lazaridis}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2896--2900}, doi = {10.2...
We describe a speech recognition system for Swiss German, a dialectal spoken language in German-speaking Switzerland. Swiss German has no standard orthography, with a significant variation in its written form. To alleviate the uncertainty associated with this variability, we automatically generate a lexicon from which ...
null
null
egorova21_interspeech
Out-of-Vocabulary Words Detection with Attention and CTC Alignments in an End-to-End ASR System
[ "Ekaterina Egorova", "Hari Krishna Vydana", "Lukáš Burget", "Jan Černocký" ]
https://www.isca-archive.org/interspeech_2021/egorova21_interspeech.html
https://www.isca-archive.org/interspeech_2021/egorova21_interspeech.pdf
10.21437/Interspeech.2021-1756
2901-2905
@inproceedings{egorova21_interspeech, title = {{Out-of-Vocabulary Words Detection with Attention and CTC Alignments in an End-to-End ASR System}}, author = {Ekaterina Egorova and Hari Krishna Vydana and Lukáš Burget and Jan Černocký}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2...
This work explores the effectiveness of detecting positions of out-of-vocabulary words (OOVs) in a decoded utterance using attention weights and CTC per-frame outputs of an end-to-end system predicting word sequences. We show that the end-to-end approach can be effective for the task of OOV detection. CTC alignments ar...
null
null
wiesner21_interspeech
Training Hybrid Models on Noisy Transliterated Transcripts for Code-Switched Speech Recognition
[ "Matthew Wiesner", "Mousmita Sarma", "Ashish Arora", "Desh Raj", "Dongji Gao", "Ruizhe Huang", "Supreet Preet", "Moris Johnson", "Zikra Iqbal", "Nagendra Goel", "Jan Trmal", "Leibny Paola García Perera", "Sanjeev Khudanpur" ]
https://www.isca-archive.org/interspeech_2021/wiesner21_interspeech.html
https://www.isca-archive.org/interspeech_2021/wiesner21_interspeech.pdf
10.21437/Interspeech.2021-2127
2906-2910
@inproceedings{wiesner21_interspeech, title = {{Training Hybrid Models on Noisy Transliterated Transcripts for Code-Switched Speech Recognition}}, author = {Matthew Wiesner and Mousmita Sarma and Ashish Arora and Desh Raj and Dongji Gao and Ruizhe Huang and Supreet Preet and Moris Johnson and Zikra Iqbal and...
In this paper, we describe the JHU-GoVivace submission for subtask 2 (code-switching task) of the Multilingual and Code-switching ASR challenges for low resource Indian languages. We built a hybrid HMM-DNN system with several improvements over the provided baseline in terms of lexical, language, and acoustic modeling. ...
null
null
xue21c_interspeech
Speech Intelligibility of Dysarthric Speech: Human Scores and Acoustic-Phonetic Features
[ "Wei Xue", "Roeland van Hout", "Fleur Boogmans", "Mario Ganzeboom", "Catia Cucchiarini", "Helmer Strik" ]
https://www.isca-archive.org/interspeech_2021/xue21c_interspeech.html
https://www.isca-archive.org/interspeech_2021/xue21c_interspeech.pdf
10.21437/Interspeech.2021-1189
2911-2915
@inproceedings{xue21c_interspeech, title = {{Speech Intelligibility of Dysarthric Speech: Human Scores and Acoustic-Phonetic Features}}, author = {Wei Xue and Roeland van Hout and Fleur Boogmans and Mario Ganzeboom and Catia Cucchiarini and Helmer Strik}, year = {2021}, booktitle = {{Interspeech 202...
We investigated speech intelligibility in dysarthric and non-dysarthric speakers as measured by two commonly used metrics, ratings through the Visual Analogue Scale (VAS) and word accuracy (AcW) through orthographic transcriptions. To gain a better understanding of how acoustic-phonetic correlates could be employed to ...
null
null
kim21i_interspeech
Analyzing Short Term Dynamic Speech Features for Understanding Behavioral Traits of Children with Autism Spectrum Disorder
[ "Young-Kyung Kim", "Rimita Lahiri", "Md. Nasir", "So Hyun Kim", "Somer Bishop", "Catherine Lord", "Shrikanth S. Narayanan" ]
https://www.isca-archive.org/interspeech_2021/kim21i_interspeech.html
https://www.isca-archive.org/interspeech_2021/kim21i_interspeech.pdf
10.21437/Interspeech.2021-2111
2916-2920
@inproceedings{kim21i_interspeech, title = {{Analyzing Short Term Dynamic Speech Features for Understanding Behavioral Traits of Children with Autism Spectrum Disorder}}, author = {Young-Kyung Kim and Rimita Lahiri and Md. Nasir and So Hyun Kim and Somer Bishop and Catherine Lord and Shrikanth S. Narayanan},...
Computational methodologies have shown promise in advancing diagnostic and intervention research in the domain of Autism Spectrum Disorder (ASD) . Prior works have investigated speech features to assess disorder severity and also to differentiate between children with and without an ASD diagnosis. In this work, we expl...
null
null
jesko21_interspeech
Vocalization Recognition of People with Profound Intellectual and Multiple Disabilities (PIMD) Using Machine Learning Algorithms
[ "Waldemar Jęśko" ]
https://www.isca-archive.org/interspeech_2021/jesko21_interspeech.html
https://www.isca-archive.org/interspeech_2021/jesko21_interspeech.pdf
10.21437/Interspeech.2021-1239
2921-2925
@inproceedings{jesko21_interspeech, title = {{Vocalization Recognition of People with Profound Intellectual and Multiple Disabilities (PIMD) Using Machine Learning Algorithms}}, author = {Waldemar Jęśko}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2921--2925}, doi = {10....
We investigate vocalization recognition for people with Profound Intellectual and Multiple Disabilities using various machine learning algorithms. The amount of training data available for people with PIMD is typically significantly limited. Due to this fact, data augmentation process was used. Various types of Machine...
null
null
fivela21_interspeech
Phonetic Complexity, Speech Accuracy and Intelligibility Assessment of Italian Dysarthric Speech
[ "Barbara Gili Fivela", "Vincenzo Sallustio", "Silvia Pede", "Danilo Patrocinio" ]
https://www.isca-archive.org/interspeech_2021/fivela21_interspeech.html
https://www.isca-archive.org/interspeech_2021/fivela21_interspeech.pdf
10.21437/Interspeech.2021-1862
2926-2930
@inproceedings{fivela21_interspeech, title = {{Phonetic Complexity, Speech Accuracy and Intelligibility Assessment of Italian Dysarthric Speech}}, author = {Barbara Gili Fivela and Vincenzo Sallustio and Silvia Pede and Danilo Patrocinio}, year = {2021}, booktitle = {{Interspeech 2021}}, pages ...
Intelligibility is the degree to which the speech of a person may be understood by a listener, and is related to functional limitation and disability. In protocols for the clinical assessment of dysarthria, intelligibility checks are included, as well as evaluations of speech accuracy, which is more directly related to...
null
null
ng21_interspeech
Detection of Consonant Errors in Disordered Speech Based on Consonant-Vowel Segment Embedding
[ "Si-Ioi Ng", "Cymie Wing-Yee Ng", "Jingyu Li", "Tan Lee" ]
https://www.isca-archive.org/interspeech_2021/ng21_interspeech.html
https://www.isca-archive.org/interspeech_2021/ng21_interspeech.pdf
10.21437/Interspeech.2021-1305
2931-2935
@inproceedings{ng21_interspeech, title = {{Detection of Consonant Errors in Disordered Speech Based on Consonant-Vowel Segment Embedding}}, author = {Si-Ioi Ng and Cymie Wing-Yee Ng and Jingyu Li and Tan Lee}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2931--2935}, doi =...
Speech sound disorder (SSD) refers to a type of developmental disorder in young children who encounter persistent difficulties in producing certain speech sounds at the expected age. Consonant errors are the major indicator of SSD in clinical assessment. Previous studies on automatic assessment of SSD revealed that det...
2106.08536
title_snapshot
hair21_interspeech
Assessing Posterior-Based Mispronunciation Detection on Field-Collected Recordings from Child Speech Therapy Sessions
[ "Adam Hair", "Guanlong Zhao", "Beena Ahmed", "Kirrie J. Ballard", "Ricardo Gutierrez-Osuna" ]
https://www.isca-archive.org/interspeech_2021/hair21_interspeech.html
https://www.isca-archive.org/interspeech_2021/hair21_interspeech.pdf
10.21437/Interspeech.2021-69
2936-2940
@inproceedings{hair21_interspeech, title = {{Assessing Posterior-Based Mispronunciation Detection on Field-Collected Recordings from Child Speech Therapy Sessions}}, author = {Adam Hair and Guanlong Zhao and Beena Ahmed and Kirrie J. Ballard and Ricardo Gutierrez-Osuna}, year = {2021}, booktitle = {...
A critical component of child speech therapy is home practice with a caregiver, who can provide feedback. However, caregivers oftentimes struggle with accurately rating speech and with perceiving pronunciation errors. One potential solution for this issue is to embed automatic mispronunciation-detection (MPD) algorithm...
null
null
mirheidari21_interspeech
Identifying Cognitive Impairment Using Sentence Representation Vectors
[ "Bahman Mirheidari", "Yilin Pan", "Daniel Blackburn", "Ronan O’Malley", "Heidi Christensen" ]
https://www.isca-archive.org/interspeech_2021/mirheidari21_interspeech.html
https://www.isca-archive.org/interspeech_2021/mirheidari21_interspeech.pdf
10.21437/Interspeech.2021-915
2941-2945
@inproceedings{mirheidari21_interspeech, title = {{Identifying Cognitive Impairment Using Sentence Representation Vectors}}, author = {Bahman Mirheidari and Yilin Pan and Daniel Blackburn and Ronan O’Malley and Heidi Christensen}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2941-...
The widely used word vectors can be extended at the sentence level to perform a wide range of natural language processing (NLP) tasks. Recently the Bidirectional Encoder Representations from Transformers (BERT) language representation achieved state-of-the-art performance for these applications. The model is trained wi...
null
null
yue21b_interspeech
Parental Spoken Scaffolding and Narrative Skills in Crowd-Sourced Storytelling Samples of Young Children
[ "Zhengjun Yue", "Jon Barker", "Heidi Christensen", "Cristina McKean", "Elaine Ashton", "Yvonne Wren", "Swapnil Gadgil", "Rebecca Bright" ]
https://www.isca-archive.org/interspeech_2021/yue21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/yue21b_interspeech.pdf
10.21437/Interspeech.2021-1297
2946-2950
@inproceedings{yue21b_interspeech, title = {{Parental Spoken Scaffolding and Narrative Skills in Crowd-Sourced Storytelling Samples of Young Children}}, author = {Zhengjun Yue and Jon Barker and Heidi Christensen and Cristina McKean and Elaine Ashton and Yvonne Wren and Swapnil Gadgil and Rebecca Bright}, ...
A novel crowdsourcing project to gather children’s storytelling based language samples using a mobile app was undertaken across the United Kingdom. Parents’ scaffolding of children’s narratives was observed in many of the samples. This study was designed to examine the relationship of scaffolding and young children’s n...
null
null
xia21_interspeech
Uncertainty-Aware COVID-19 Detection from Imbalanced Sound Data
[ "Tong Xia", "Jing Han", "Lorena Qendro", "Ting Dang", "Cecilia Mascolo" ]
https://www.isca-archive.org/interspeech_2021/xia21_interspeech.html
https://www.isca-archive.org/interspeech_2021/xia21_interspeech.pdf
10.21437/Interspeech.2021-1320
2951-2955
@inproceedings{xia21_interspeech, title = {{Uncertainty-Aware COVID-19 Detection from Imbalanced Sound Data}}, author = {Tong Xia and Jing Han and Lorena Qendro and Ting Dang and Cecilia Mascolo}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2951--2955}, doi = {10.21437/In...
Recently, sound-based COVID-19 detection studies have shown great promise to achieve scalable and prompt digital pre-screening. However, there are still two unsolved issues hindering the practice. First, collected datasets for model training are often imbalanced, with a considerably smaller proportion of users tested p...
2104.02005
title_snapshot
wang21u_interspeech
Unsupervised Domain Adaptation for Dysarthric Speech Detection via Domain Adversarial Training and Mutual Information Minimization
[ "Disong Wang", "Liqun Deng", "Yu Ting Yeung", "Xiao Chen", "Xunying Liu", "Helen Meng" ]
https://www.isca-archive.org/interspeech_2021/wang21u_interspeech.html
https://www.isca-archive.org/interspeech_2021/wang21u_interspeech.pdf
10.21437/Interspeech.2021-2139
2956-2960
@inproceedings{wang21u_interspeech, title = {{Unsupervised Domain Adaptation for Dysarthric Speech Detection via Domain Adversarial Training and Mutual Information Minimization}}, author = {Disong Wang and Liqun Deng and Yu Ting Yeung and Xiao Chen and Xunying Liu and Helen Meng}, year = {2021}, boo...
Dysarthric speech detection (DSD) systems aim to detect characteristics of the neuromotor disorder from speech. Such systems are particularly susceptible to domain mismatch where the training and testing data come from the source and target domains respectively, but the two domains may differ in terms of speech stimuli...
2106.10127
title_snapshot
bhattacharjee21_interspeech
Source and Vocal Tract Cues for Speech-Based Classification of Patients with Parkinson’s Disease and Healthy Subjects
[ "Tanuka Bhattacharjee", "Jhansi Mallela", "Yamini Belur", "Nalini Atchayaram", "Ravi Yadav", "Pradeep Reddy", "Dipanjan Gope", "Prasanta Kumar Ghosh" ]
https://www.isca-archive.org/interspeech_2021/bhattacharjee21_interspeech.html
https://www.isca-archive.org/interspeech_2021/bhattacharjee21_interspeech.pdf
10.21437/Interspeech.2021-2008
2961-2965
@inproceedings{bhattacharjee21_interspeech, title = {{Source and Vocal Tract Cues for Speech-Based Classification of Patients with Parkinson’s Disease and Healthy Subjects}}, author = {Tanuka Bhattacharjee and Jhansi Mallela and Yamini Belur and Nalini Atchayaram and Ravi Yadav and Pradeep Reddy and Dipanjan...
Parkinson’s disease (PD) affects both source and vocal tract components of speech. Various speech cues explored in literature for automatic classification of individuals with PD and healthy controls (HC) implicitly carry information about both these components. This work explicitly analyzes the contribution of source a...
null
null
haulcy21_interspeech
CLAC: A Speech Corpus of Healthy English Speakers
[ "R’mani Haulcy", "James Glass" ]
https://www.isca-archive.org/interspeech_2021/haulcy21_interspeech.html
https://www.isca-archive.org/interspeech_2021/haulcy21_interspeech.pdf
10.21437/Interspeech.2021-1810
2966-2970
@inproceedings{haulcy21_interspeech, title = {{CLAC: A Speech Corpus of Healthy English Speakers}}, author = {R’mani Haulcy and James Glass}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2966--2970}, doi = {10.21437/Interspeech.2021-1810}, issn = {2958-1796}, }
This paper introduces the Crowdsourced Language Assessment Corpus (CLAC), a speech corpus consisting of audio recordings and automatically-generated transcripts for several speech and language tasks, as well as metadata for each of the speakers. The CLAC was created to provide the community with a collection of audio s...
null
null
nortje21_interspeech
Direct Multimodal Few-Shot Learning of Speech and Images
[ "Leanne Nortje", "Herman Kamper" ]
https://www.isca-archive.org/interspeech_2021/nortje21_interspeech.html
https://www.isca-archive.org/interspeech_2021/nortje21_interspeech.pdf
10.21437/Interspeech.2021-49
2971-2975
@inproceedings{nortje21_interspeech, title = {{Direct Multimodal Few-Shot Learning of Speech and Images}}, author = {Leanne Nortje and Herman Kamper}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2971--2975}, doi = {10.21437/Interspeech.2021-49}, issn = {2958-1796},...
We propose direct multimodal few-shot models that learn a shared embedding space of spoken words and images from only a few paired examples. Imagine an agent is shown an image along with a spoken word describing the object in the picture, e.g. pen, book and eraser . After observing a few paired examples of each class, ...
2012.05680
title_snapshot
sanabria21_interspeech
Talk, Don’t Write: A Study of Direct Speech-Based Image Retrieval
[ "Ramon Sanabria", "Austin Waters", "Jason Baldridge" ]
https://www.isca-archive.org/interspeech_2021/sanabria21_interspeech.html
https://www.isca-archive.org/interspeech_2021/sanabria21_interspeech.pdf
10.21437/Interspeech.2021-96
2976-2980
@inproceedings{sanabria21_interspeech, title = {{Talk, Don’t Write: A Study of Direct Speech-Based Image Retrieval}}, author = {Ramon Sanabria and Austin Waters and Jason Baldridge}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2976--2980}, doi = {10.21437/Interspeech.2021...
Speech-based image retrieval has been studied as a proxy for joint representation learning, usually without emphasis on retrieval itself. As such, it is unclear how well speech-based retrieval can work in practice — both in an absolute sense and versus alternative strategies that combine automatic speech recognition (A...
2104.01894
title_snapshot
zhao21b_interspeech
A Fast Discrete Two-Step Learning Hashing for Scalable Cross-Modal Retrieval
[ "Huan Zhao", "Kaili Ma" ]
https://www.isca-archive.org/interspeech_2021/zhao21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/zhao21b_interspeech.pdf
10.21437/Interspeech.2021-287
2981-2985
@inproceedings{zhao21b_interspeech, title = {{A Fast Discrete Two-Step Learning Hashing for Scalable Cross-Modal Retrieval}}, author = {Huan Zhao and Kaili Ma}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2981--2985}, doi = {10.21437/Interspeech.2021-287}, issn = {...
Recently, some cross-modal hashing methods are proposed to search data for different modality effectively. Hashing has received wide attention because of its low storage and high efficiency. Hashing-based methods project the data instances from different modalities into a Hamming space to learn hash codes for retrieval...
null
null
wang21v_interspeech
Cross-Modal Knowledge Distillation Method for Automatic Cued Speech Recognition
[ "Jianrong Wang", "Ziyue Tang", "Xuewei Li", "Mei Yu", "Qiang Fang", "Li Liu" ]
https://www.isca-archive.org/interspeech_2021/wang21v_interspeech.html
https://www.isca-archive.org/interspeech_2021/wang21v_interspeech.pdf
10.21437/Interspeech.2021-432
2986-2990
@inproceedings{wang21v_interspeech, title = {{Cross-Modal Knowledge Distillation Method for Automatic Cued Speech Recognition}}, author = {Jianrong Wang and Ziyue Tang and Xuewei Li and Mei Yu and Qiang Fang and Li Liu}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2986--2990}, ...
Cued Speech (CS) is a visual communication system for the deaf or hearing impaired people. It combines lip movements with hand cues to obtain a complete phonetic repertoire. Current deep learning based methods on automatic CS recognition suffer from a common problem, which is the data scarcity. Until now, there are onl...
2106.13686
title_snapshot
olaleye21_interspeech
Attention-Based Keyword Localisation in Speech Using Visual Grounding
[ "Kayode Olaleye", "Herman Kamper" ]
https://www.isca-archive.org/interspeech_2021/olaleye21_interspeech.html
https://www.isca-archive.org/interspeech_2021/olaleye21_interspeech.pdf
10.21437/Interspeech.2021-435
2991-2995
@inproceedings{olaleye21_interspeech, title = {{Attention-Based Keyword Localisation in Speech Using Visual Grounding}}, author = {Kayode Olaleye and Herman Kamper}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2991--2995}, doi = {10.21437/Interspeech.2021-435}, issn ...
Visually grounded speech models learn from images paired with spoken captions. By tagging images with soft text labels using a trained visual classifier with a fixed vocabulary, previous work has shown that it is possible to train a model that can detect whether a particular text keyword occurs in speech utterances or ...
2106.08859
title_snapshot
khorrami21_interspeech
Evaluation of Audio-Visual Alignments in Visually Grounded Speech Models
[ "Khazar Khorrami", "Okko Räsänen" ]
https://www.isca-archive.org/interspeech_2021/khorrami21_interspeech.html
https://www.isca-archive.org/interspeech_2021/khorrami21_interspeech.pdf
10.21437/Interspeech.2021-496
2996-3000
@inproceedings{khorrami21_interspeech, title = {{Evaluation of Audio-Visual Alignments in Visually Grounded Speech Models}}, author = {Khazar Khorrami and Okko Räsänen}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {2996--3000}, doi = {10.21437/Interspeech.2021-496}, issn...
Systems that can find correspondences between multiple modalities, such as between speech and images, have great potential to solve different recognition and data analysis tasks in an unsupervised manner. This work studies multimodal learning in the context of visually grounded speech (VGS) models, and focuses on their...
2108.02562
title_snapshot
chen21k_interspeech
Automatic Lip-Reading with Hierarchical Pyramidal Convolution and Self-Attention for Image Sequences with No Word Boundaries
[ "Hang Chen", "Jun Du", "Yu Hu", "Li-Rong Dai", "Bao-Cai Yin", "Chin-Hui Lee" ]
https://www.isca-archive.org/interspeech_2021/chen21k_interspeech.html
https://www.isca-archive.org/interspeech_2021/chen21k_interspeech.pdf
10.21437/Interspeech.2021-723
3001-3005
@inproceedings{chen21k_interspeech, title = {{Automatic Lip-Reading with Hierarchical Pyramidal Convolution and Self-Attention for Image Sequences with No Word Boundaries}}, author = {Hang Chen and Jun Du and Yu Hu and Li-Rong Dai and Bao-Cai Yin and Chin-Hui Lee}, year = {2021}, booktitle = {{Inter...
In this paper, we propose a novel deep learning architecture for improving word-level lip-reading. We first incorporate multi-scale processing into spatial feature extraction for lip-reading using hierarchical pyramidal convolution (HPConv) and self-attention. Specifically, HPConv is proposed to replace the conventiona...
2012.14360
title_judge
rouditchenko21b_interspeech
Cascaded Multilingual Audio-Visual Learning from Videos
[ "Andrew Rouditchenko", "Angie Boggust", "David Harwath", "Samuel Thomas", "Hilde Kuehne", "Brian Chen", "Rameswar Panda", "Rogerio Feris", "Brian Kingsbury", "Michael Picheny", "James Glass" ]
https://www.isca-archive.org/interspeech_2021/rouditchenko21b_interspeech.html
https://www.isca-archive.org/interspeech_2021/rouditchenko21b_interspeech.pdf
10.21437/Interspeech.2021-1352
3006-3010
@inproceedings{rouditchenko21b_interspeech, title = {{Cascaded Multilingual Audio-Visual Learning from Videos}}, author = {Andrew Rouditchenko and Angie Boggust and David Harwath and Samuel Thomas and Hilde Kuehne and Brian Chen and Rameswar Panda and Rogerio Feris and Brian Kingsbury and Michael Picheny and...
In this paper, we explore self-supervised audio-visual models that learn from instructional videos. Prior work has shown that these models can relate spoken words and sounds to visual content after training on a large-scale dataset of videos, but they were only trained and evaluated on videos in English. To learn multi...
2111.04823
title_snapshot
ma21c_interspeech
LiRA: Learning Visual Speech Representations from Audio Through Self-Supervision
[ "Pingchuan Ma", "Rodrigo Mira", "Stavros Petridis", "Björn W. Schuller", "Maja Pantic" ]
https://www.isca-archive.org/interspeech_2021/ma21c_interspeech.html
https://www.isca-archive.org/interspeech_2021/ma21c_interspeech.pdf
10.21437/Interspeech.2021-1360
3011-3015
@inproceedings{ma21c_interspeech, title = {{LiRA: Learning Visual Speech Representations from Audio Through Self-Supervision}}, author = {Pingchuan Ma and Rodrigo Mira and Stavros Petridis and Björn W. Schuller and Maja Pantic}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {3011--3...
The large amount of audiovisual content being shared online today has drawn substantial attention to the prospect of audio-visual self-supervised learning. Recent works have focused on each of these modalities separately, while others have attempted to model both simultaneously in a cross-modal fashion. However, compar...
2106.09171
title_snapshot
rose21_interspeech
End-to-End Audio-Visual Speech Recognition for Overlapping Speech
[ "Richard Rose", "Olivier Siohan", "Anshuman Tripathi", "Otavio Braga" ]
https://www.isca-archive.org/interspeech_2021/rose21_interspeech.html
https://www.isca-archive.org/interspeech_2021/rose21_interspeech.pdf
10.21437/Interspeech.2021-1621
3016-3020
@inproceedings{rose21_interspeech, title = {{End-to-End Audio-Visual Speech Recognition for Overlapping Speech}}, author = {Richard Rose and Olivier Siohan and Anshuman Tripathi and Otavio Braga}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {3016--3020}, doi = {10.21437/In...
This paper investigates an end-to-end audio-visual (A/V) modeling approach for transcribing utterances in scenarios where there are overlapping speech utterances from multiple talkers. It assumes that overlapping audio signals and video signals in the form of mouth-tracks aligned with speech are available for overlappi...
null
null
wu21e_interspeech
Audio-Visual Multi-Talker Speech Recognition in a Cocktail Party
[ "Yifei Wu", "Chenda Li", "Song Yang", "Zhongqin Wu", "Yanmin Qian" ]
https://www.isca-archive.org/interspeech_2021/wu21e_interspeech.html
https://www.isca-archive.org/interspeech_2021/wu21e_interspeech.pdf
10.21437/Interspeech.2021-2128
3021-3025
@inproceedings{wu21e_interspeech, title = {{Audio-Visual Multi-Talker Speech Recognition in a Cocktail Party}}, author = {Yifei Wu and Chenda Li and Song Yang and Zhongqin Wu and Yanmin Qian}, year = {2021}, booktitle = {{Interspeech 2021}}, pages = {3021--3025}, doi = {10.21437/Inters...
Speech from microphones is vulnerable in a complex acoustic environment due to noise and reverberation, while the cameras are not. Thus, utilizing the visual modality in the “cocktail party” scenario with multi-talkers has become a promising and popular approach. In this paper, we have explored the incorporating of vis...
null
null
chen21l_interspeech
Ultra Fast Speech Separation Model with Teacher Student Learning
[ "Sanyuan Chen", "Yu Wu", "Zhuo Chen", "Jian Wu", "Takuya Yoshioka", "Shujie Liu", "Jinyu Li", "Xiangzhan Yu" ]
https://www.isca-archive.org/interspeech_2021/chen21l_interspeech.html
https://www.isca-archive.org/interspeech_2021/chen21l_interspeech.pdf
10.21437/Interspeech.2021-142
3026-3030
@inproceedings{chen21l_interspeech, title = {{Ultra Fast Speech Separation Model with Teacher Student Learning}}, author = {Sanyuan Chen and Yu Wu and Zhuo Chen and Jian Wu and Takuya Yoshioka and Shujie Liu and Jinyu Li and Xiangzhan Yu}, year = {2021}, booktitle = {{Interspeech 2021}}, pages ...
Transformer has been successfully applied to speech separation recently with its strong long-dependency modeling capacity using a self-attention mechanism. However, Transformer tends to have heavy run-time costs due to the deep encoder layers, which hinders its deployment on edge devices. A small Transformer model with...
2204.12777
title_snapshot