The DOI matches its Crossref record; 78 of the 86 references it cites were found in a registry or an index.
Toward Human Parity in Conversational Speech Recognition
Abstract
Conversational speech recognition has served as a flagship speech recognition task since the release of the Switchboard corpus in the 1990s. In this paper, we measure a human error rate on the widely used NIST 2000 test set for commercial bulk transcription. The error rate of professional transcribers is 5.9% for the Switchboard portion of the data, in which newly acquainted pairs of people discuss an assigned topic, and 11.3% for the CallHome portion, where friends and family members have open-ended conversations. In both cases, our automated system edges past the human benchmark, achieving error rates of 5.8% and 11.0%, respectively. The key to our system's performance is the use of various convolutional and long-short-term memory acoustic model architectures, combined with a novel spatial smoothing method and lattice-free discriminative acoustic training, multiple recurrent neural network language modeling approaches, and a systematic use of system combination. Comparing frequent errors in our human and machine transcripts, we find them to be remarkably similar, and highly correlated as a function of the speaker. Human subjects find it very difficult to tell which errorful transcriptions come from humans. Overall, this suggests that, given sufficient matched training data, conversational speech transcription engines are approximating human parity in both quantitative and qualitative terms.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
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?
In the citing paper
“In ASR, a 5% WER, over longer segments of text, is standard for professional-level transcriptions141 and 25% or less is considered acceptable142.”
From The speech neuroprosthesis (2024) · cites this paper as [141]
In the citing paper
“Spontaneous, conversational speech is not linguistically well-formed.21”
From Challenges of developing a digital scribe to reduce clinical documentation burden (2019) · cites this paper as [21]
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 1×
“The second network is modeled on the ResNet architecture [50], which adds highway connections, i.”
BackgroundVery Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148 · cited 1×
“The first is the VGG architecture of [49].”
BackgroundHochreiter, S., & Schmidhuber, J. (1997) · Neural Computation · doi:10.1162/neco.1997.9.8.1735 · Cited by 102,318 · cited 1×
“While the basic structures have been well known for a long period [21]–[27], they have only recently started to dominate the field as”
BackgroundAdam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946 · cited 1×
“We use the “FsAdaGrad” learning algorithm, which is an implementation of Adam [77] in CNTK.”
MethodBatch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Ioffe, S., & Szegedy, C. (2015) · arXiv · doi:10.48550/arxiv.1502.03167 · Cited by 27,818
Sequence to Sequence Learning with Neural Networks
Sutskever, I., Vinyals, O., & Le, Q. V. (2014) · arXiv · doi:10.48550/arxiv.1409.3215 · Cited by 17,766 · cited 1×
“[42] ensemble learning has now also become a feature in neural models [35], [43], [44], to improve robustness by reducing bias and variance.”
BackgroundMastering the game of Go with deep neural networks and tree search
Silver, D., Huang, A., Maddison, C. J., Guez, A., Sifre, L., van den Driessche, G., Schrittwieser, J., Antonoglou, I., Panneershelvam, V., Lanctot, M., et al. (2016) · Nature · doi:10.1038/nature16961 · Cited by 15,612
Backpropagation Applied to Handwritten Zip Code Recognition
LeCun, Y., Boser, B., Denker, J. S., Henderson, D., Howard, R. E., Hubbard, W., & Jackel, L. D. (1989) · Neural Computation · doi:10.1162/neco.1989.1.4.541 · Cited by 11,891 · cited 1×
“While the basic structures have been well known for a long period [21], [22], [23], [24], [25], [26], [27], they have only recently started to dominate the field as the best models for”
BackgroundFramewise phoneme classification with bidirectional LSTM and other neural network architectures
Graves, A., & Schmidhuber, J. (2005) · Neural Networks · doi:10.1016/j.neunet.2005.06.042 · Cited by 5,839
Recurrent neural network based language model
Mikolov, T., Karafiát, M., Burget, L., Černocký, J., & Khudanpur, S. (2010) · Interspeech 2010 · doi:10.21437/interspeech.2010-343 · Cited by 5,390 · cited 2×
“This is the case for both acoustic modeling [28]–[33] and language modeling [34]–[37].”
Background
Research works printed but not linked: 3
- 10.1007/3-540-33486-6_6 · doi:10.1007/3-540-33486-6_6 · Paper page
- ljolje, 0, The AT&T 2001 LVCSR system, Proc NIST LVCSR Workshop
- graff, 0, The 1996 broadcast news speech and language-model corpus, Proc DARPA Spoken Language Systems Technology Workshop, 11
Cited by
Artificial intelligence in radiology
Hosny, A., Parmar, C., Quackenbush, J., Schwartz, L. H., & Aerts, H. J. W. L. (2018) · Nature Reviews Cancer · doi:10.1038/s41568-018-0016-5 · Cited by 3,803
“Currently , we are witnessing narrow task-specific AI applications that are able to match and occasionally surpass human intelligence 4–6 , 9 . are vital.”
On calibration of modern neural networks
Guo, C., Pleiss, G., Sun, Y., & Weinberger, K. Q. (2017) · arXiv · doi:10.48550/arxiv.1706.04599 · Cited by 3,435
Opportunities and obstacles for deep learning in biology and medicine
Ching, T., Himmelstein, D. S., Beaulieu-Jones, B. K., Kalinin, A. A., Do, B. T., Way, G. P., Ferrero, E., Agapow, P.-M., Zietz, M., Hoffman, M. M., et al. (2018) · Journal of The Royal Society Interface · doi:10.1098/rsif.2017.0387 · Cited by 2,310
DeepXplore: Automated Whitebox Testing of Deep Learning Systems
Pei, K., Cao, Y., Yang, J., & Jana, S. (2017) · Proceedings of the 26th Symposium on Operating Systems Principles · doi:10.1145/3132747.3132785 · Cited by 1,217
Targeted Backdoor Attacks on Deep Learning Systems Using Data Poisoning
Chen, X., Liu, C., Li, B., Lu, K., & Song, D. (2017) · arXiv · doi:10.48550/arxiv.1712.05526 · Cited by 1,039
Metadata
- Authors
- Xiong, W., Droppo, J., Huang, X., Seide, F., Seltzer, M. L., Stolcke, A., Yu, D., & Zweig, G.
- Affiliations
- Microsoft (United States) (US)
- Venue
- IEEE/ACM Transactions on Audio, Speech, and Language Processing
- Published
- 2017-12
- Volume
- 25
- Issue
- 12
- Pages
- 2410-2423
- ISSN
- 2329-9290 (ISSN-L)
- 2329-9304
- OpenAlex
- W2759071281
- Cited by
- 245
- References
- 86
- Citations by year
- 2026: 11 · 2025: 12 · 2024: 23 · 2023: 27 · 2022: 32 · 2021: 44 · 2020: 39 · 2019: 37 · 2018: 17 · 2017: 3
- FWCI
- 15.67
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- IEEE/ACM Transactions on Audio Speech and Language Processing (publishedVersion)
- Funders
- University of Cambridge
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
- Physical Sciences › Computer Science › Signal Processing › Speech and Audio Processing
- Physical Sciences › Computer Science › Artificial Intelligence › Speech and dialogue systems
- Keywords
- conversational speech recognition, human error rates, Switchboard corpus, CallHome corpus, acoustic modeling, convolutional neural networks, long short-term memory, spatial smoothing, system combination, error analysis, speaker differences
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Toward Human Parity in Conversational Speech Recognition. https://app.cochat.ai/papers/10.1109/taslp.2017.2756440
- The paper
- Xiong, W., Droppo, J., Huang, X., Seide, F., Seltzer, M. L., Stolcke, A., Yu, D., & Zweig, G. (2017). Toward Human Parity in Conversational Speech Recognition. IEEE/ACM Transactions on Audio, Speech, and Language Processing. https://doi.org/10.1109/taslp.2017.2756440
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.