The DOI matches its Crossref record; 23 of the 23 references it cites were found in a registry or an index.
SEP-28k: A Dataset for Stuttering Event Detection from Podcasts with People Who Stutter
Open accessAbstract
The ability to automatically detect stuttering events in speech could help speech pathologists track an individual’s fluency over time or help improve speech recognition systems for people with atypical speech patterns. Despite increasing interest in this area, existing public datasets are too small to build generalizable dysfluency detection systems and lack sufficient annotations. In this work, we introduce Stuttering Events in Podcasts (SEP-28k), a dataset containing over 28k clips labeled with five event types including blocks, prolongations, sound repetitions, word repetitions, and interjections. Audio comes from public podcasts largely consisting of people who stutter interviewing other people who stutter. We benchmark a set of acoustic models on SEP-28k and the public FluencyBank dataset and highlight how simply increasing the amount of training data improves relative detection performance by 28% and 24% F1 on each. Annotations from over 32k clips across both datasets will be publicly released.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Similar limitations are common in accessibility datasets, such as sign languages [12, 27] and stuttered speech [58], which rarely involve affected communities in the annotation process.”
From Disability-First AI Dataset Annotation: Co-designing Stuttered Speech Annotation Guidelines with People Who Stutter (2026) · cites this paper as [58]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“In contrast to conventional annotation guidelines that focus primarily on audible speech [58, 83]”
From Disability-First AI Dataset Annotation: Co-designing Stuttered Speech Annotation Guidelines with People Who Stutter (2026) · cites this paper as [58]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“However, access to these datasets is often restricted, as they typically involve sensitive or proprietary data, except when sourced from public domains such as podcasts [58].”
From Disability-First AI Dataset Annotation: Co-designing Stuttered Speech Annotation Guidelines with People Who Stutter (2026) · cites this paper as [58]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Sep-28k consists of over 28k three-second audio clips curated from public podcasts featuring PWS; however, the annotators' backgrounds remain unclear.”
From Disability-First AI Dataset Annotation: Co-designing Stuttered Speech Annotation Guidelines with People Who Stutter (2026) · cites this paper as [58]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“as the largest publicly available English stuttering dataset, Sep-28k does not have transcriptions and provides only clip-level annotations rather than word-/syllablelevel labels [58].”
From Disability-First AI Dataset Annotation: Co-designing Stuttered Speech Annotation Guidelines with People Who Stutter (2026) · cites this paper as [58]
Experimental — accuracy not yet measured
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“We focused on Sep-28k due to its widespread use 4 , its public availability of annotations produced by PWS [92] and non-PWS [58], and the reported discrepancy between these two versions [92].”
From Disability-First AI Dataset Annotation: Co-designing Stuttered Speech Annotation Guidelines with People Who Stutter (2026) · cites this paper as [58]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“We adopted the five stuttering events used in Sep-28k as our initial ontology, as these events represent the core disfluency types identified across prior work [32, 58, 63, 83].”
From Disability-First AI Dataset Annotation: Co-designing Stuttered Speech Annotation Guidelines with People Who Stutter (2026) · cites this paper as [58]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“Sep-28k reported an inter-rater agreement of 0.11 for identifying speech prolongations, 0.25 for blocks, and 0.39 for determining whether a clip contains disfluencies [58].”
From Disability-First AI Dataset Annotation: Co-designing Stuttered Speech Annotation Guidelines with People Who Stutter (2026) · cites this paper as [58]
The cited source says what the claim attributes to it.
In the citing paper
“Similar label inconsistencies have been observed in datasets of sign languages [12] and stuttered speech [58, 92].”
From Disability-First AI Dataset Annotation: Co-designing Stuttered Speech Annotation Guidelines with People Who Stutter (2026) · cites this paper as [58]
The cited source says what the claim attributes to it.
In the citing paper
“As Sep-28k divides audio into strictly 3-second clips [58], it is often difficult to determine stuttering events without access to the conversational flow and context.”
From Disability-First AI Dataset Annotation: Co-designing Stuttered Speech Annotation Guidelines with People Who Stutter (2026) · cites this paper as [58]
The cited source says what the claim attributes to it.
In the citing paper
“Our guidelines differ from SLP frameworks developed to measure fluency and stuttering severity [83, 109], as well as from earlier labeling frameworks designed primarily to detect stuttering events [58].”
From Disability-First AI Dataset Annotation: Co-designing Stuttered Speech Annotation Guidelines with People Who Stutter (2026) · cites this paper as [58]
The check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“people who stutter (PWS) are rarely involved to address such complexities [6, 58, 83, 102].”
From Disability-First AI Dataset Annotation: Co-designing Stuttered Speech Annotation Guidelines with People Who Stutter (2026) · cites this paper as [58]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Focal Loss for Dense Object Detection
Lin, T.-Y., Goyal, P., Girshick, R., He, K., & Dollar, P. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.324 · Cited by 30,581
“Librispeech: an ASR corpus based on public domain audio books,”
Panayotov, V., Chen, G., Povey, D., & Khudanpur, S. (2015) · 2015 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2015.7178964 · Cited by 6,072
The University College London Archive of Stuttered Speech (UCLASS)
Howell, P., Davis, S., & Bartrip, J. (2009) · Journal of Speech Language and Hearing Research · doi:10.1044/1092-4388(2009/07-0129 · Cited by 127
Fluency Bank: A new resource for fluency research and practice
Bernstein Ratner, N., & MacWhinney, B. (2018) · Journal of Fluency Disorders · doi:10.1016/j.jfludis.2018.03.002 · Cited by 124
Sequence-to-Sequence Speech Recognition with Time-Depth Separable Convolutions
Hannun, A., Lee, A., Xu, Q., & Collobert, R. (2019) · Interspeech (USB) · doi:10.21437/interspeech.2019-2460 · Cited by 103
Kourkounakis, T., Hajavi, A., & Etemad, A. (2020) · ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp40776.2020.9053893 · Cited by 87
Intelligent Processing of Stuttered Speech
Czyżewski, A., Kaczmarek, A., & Kostek, B. (2003) · Journal of Intelligent Information Systems · doi:10.1023/a:1024710532716 · Cited by 65
Retrieving Tract Variables From Acoustics: A Comparison of Different Machine Learning Strategies
Mitra, V., Nam, H., Espy-Wilson, C., Saltzman, E. L., & Goldstein, L. M. (2010) · IEEE Journal of Selected Topics in Signal Processing · doi:10.1109/jstsp.2010.2076013 · Cited by 56
Self-citationKane, S. K., Guo, A., & Morris, M. R. (2020) · International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS) · doi:10.1145/3373625.3416990 · Cited by 36
Time-Interval Measurement of Stuttering
Ingham, R. J., Cordes, A. K., & Finn, P. (1993) · Journal of Speech Language and Hearing Research · doi:10.1044/jshr.3606.1168 · Cited by 35
A Lightly Supervised Approach to Detect Stuttering in Children's Speech
Alharbi, S., Hasan, M., Simons, A. J. H., Brumfitt, S. M., & Green, P. (2018) · Interspeech (USB) · doi:10.21437/interspeech.2018-2155 · Cited by 35
Brewer, R. N., Findlater, L., Kaye, J. '., Lasecki, W., Munteanu, C., & Weber, A. (2018) · Companion of the 2018 ACM Conference on Computer Supported Cooperative Work and Social Computing · doi:10.1145/3272973.3273006 · Cited by 32
Speech diversity and speech interfaces: considering an inclusive future through stammering
Clark, L., Cowan, B. R., Roper, A., Lindsay, S., & Sheers, O. (2020) · CIU · doi:10.1145/3405755.3406139 · Cited by 30
Speech Recognition and Correction of a Stuttered Speech
Dash, A., Subramani, N., Manjunath, T., Yaragarala, V., & Tripathi, S. (2018) · International Conference on Advances in Computing, Communications and Informatics (ICACCI) · doi:10.1109/icacci.2018.8554455 · Cited by 28
Automatic Correction of Stutter in Disfluent Speech
N, A. K., Karthik, S., Kamalnath, D., Chanda, P., & Tripathi, S. (2020) · Procedia Computer Science · doi:10.1016/j.procs.2020.04.146 · Cited by 24
Gaussian Mixture Model Based Classification of Stuttering Dysfluencies
Mahesha, P., & Vinod, D. S. (2015) · Journal of Intelligent Systems · doi:10.1515/jisys-2014-0140 · Cited by 21
Using Clinician Annotations to Improve Automatic Speech Recognition of Stuttered Speech
Heeman, P. A., Lunsford, R., McMillin, A., & Yaruss, J. S. (2016) · Interspeech (USB) · doi:10.21437/interspeech.2016-1388 · Cited by 19
Towards Automated Assessment of Stuttering and Stuttering Therapy
Bayerl, S. P., Hönig, F., Reister, J., & Riedhammer, K. (2020) · Lecture notes in computer science · doi:10.1007/978-3-030-58323-1_42 · Cited by 17
Leveraging Acoustic Cues and Paralinguistic Embeddings to Detect Expression from Voice
Mitra, V., Booker, S., Marchi, E., Farrar, D. S., Peitz, U. D., Cheng, B., Teves, E. A., Mehta, A., & Naik, D. K. (2019) · Interspeech (USB) · doi:10.21437/interspeech.2019-2998 · Cited by 17
Self-citationIdentification of primary and collateral tracks in stuttered speechPreprint
Riad, R., Bachoud‐Lévi, A., Rudzicz, F., & Emmanuel Dupoux, A. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2003.01018 · Cited by 11
Event‐ and interval‐based measurement of stuttering: a review
Valente, A. R., Jesus, L. M. T., Hall, A. O., & Leahy, M. M. (2014) · International Journal of Language & Communication Disorders · doi:10.1111/1460-6984.12113 · Cited by 10
Automatic recognition of children's read speech for stuttering application
Alharbi, S., Simons, A. J. H., Brumfitt, S. M., & Green, P. (2017) · Workshop on Child, Computer and Interaction (WOCCI) · doi:10.21437/wocci.2017-1 · Cited by 10
Noise Robust Acoustic to Articulatory Speech Inversion
Seneviratne, N., Sivaraman, G., Mitra, V., & Espy-Wilson, C. (2018) · Interspeech (USB) · doi:10.21437/interspeech.2018-1509 · Cited by 9
Self-citation
Research works printed but not linked: 4
- devis, 2009, The UCLASS archive of stuttered speech, J Speech Lang Hear Res
- soundararajan, 2020, Stammering accessibility and testing for voice assistants devices, Personal blog
- van riper, 1983, The nature of stuttering, Applied Psycholinguistics
- corcoran, 2018, When alexa can’t understand you, Slate
Cited by
Sorted by citations
Audio Deepfake Detection: What Has Been Achieved and What Lies Ahead
Zhang, B., Cui, H., Nguyen, V., & Whitty, M. (2025) · Sensors · doi:10.3390/s25071989 · Cited by 72
Automation of Language Sample Analysis
Liu, H., MacWhinney, B., Fromm, D., & Lanzi, A. (2023) · Journal of Speech, Language, and Hearing Research · doi:10.1044/2023_jslhr-22-00642 · Cited by 59
Lea, C., Huang, Z., Narain, J., Tooley, L., Yee, D., Tran, D. T., Georgiou, P., Bigham, J. P., & Findlater, L. (2023) · Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems · doi:10.1145/3544548.3581224 · Cited by 35
Self-citationDysfluency Classification in Stuttered Speech Using Deep Learning for Real-Time Applications
Jouaiti, M., & Dautenhahn, K. (2022) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp43922.2022.9746638 · Cited by 35
Systematic Review of Machine Learning Approaches for Detecting Developmental Stuttering
Barrett, L., Hu, J., & Howell, P. (2022) · IEEE/ACM Transactions on Audio Speech and Language Processing · doi:10.1109/taslp.2022.3155295 · Cited by 35
Metadata
- Authors
- Lea, C., Mitra, V., Joshi, A., Kajarekar, S., & Bigham, J. P.
- Affiliations
- Apple (United States) (US)
- Venue
- ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
- Publisher
- IEEE
- Published
- 2021-06-06
- Pages
- 6798-6802
- ISSN
- 1520-6149 (ISSN-L)
- OpenAlex
- W3132874366
- Cited by
- 113
- References
- 23
- Citations by year
- 2026: 13 · 2025: 29 · 2024: 33 · 2023: 21 · 2022: 15 · 2021: 2
- FWCI
- 140.62
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- Locations
- IEEE International Conference on Acoustics Speech and Signal Processing (publishedVersion)
- arXiv (Cornell University) (submittedVersion, public-domain)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- IEEE RESOURCE CENTERS
- Topics
- Social Sciences › Psychology › Clinical Psychology › Stuttering Research and Treatment
- Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
- Physical Sciences › Computer Science › Signal Processing › Speech and Audio Processing
- Keywords
- stuttering, podcasts, speech annotation, prolongation, sound repetition, word repetition, interjections, training data size
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
