Toward Human Parity in Conversational Speech Recognition
The DOI matches its Crossref record; 78 of the 86 references it cites were found in a registry or an index.
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].”
BackgroundA Learning Algorithm for Continually Running Fully Recurrent Neural Networks
Williams, R. J., & Zipser, D. (1989) · Neural Computation · doi:10.1162/neco.1989.1.2.270 · Cited by 4,369 · 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”
BackgroundSRILM - an extensible language modeling toolkit
Stolcke, A. (2002) · 7th International Conference on Spoken Language Processing (ICSLP 2002) · doi:10.21437/icslp.2002-303 · Cited by 4,326 · cited 1×
“We use an N-gram language model trained and pruned with the SRILM toolkit [66].”
MethodSelf-citationConvolutional networks for images, speech, and time series
LeCun, Y., & Bengio, Y. (1998) · HAL (Le Centre pour la Communication Scientifique Directe) · doi:10.5555/303568.303704 · Cited by 4,231
Front-End Factor Analysis for Speaker Verification
Dehak, N., Kenny, P. J., Dehak, R., Dumouchel, P., & Ouellet, P. (2011) · IEEE Transactions on Audio, Speech, and Language Processing · doi:10.1109/tasl.2010.2064307 · Cited by 3,549
Noise-contrastive estimation: A new estimation principle for unnormalized statistical models
Gutmann, M. U., & Hyvärinen, A. (2010) · Edinburgh Research Explorer (University of Edinburgh) · Cited by 2,992
Long short-term memory recurrent neural network architectures for large scale acoustic modeling
Sak, H., Senior, A., & Beaufays, F. (2014) · Interspeech 2014 · doi:10.21437/interspeech.2014-80 · Cited by 2,973 · cited 2×
“This is the case for both acoustic modeling [28]–[33] and language modeling [34]–[37].”
BackgroundLinguistic Regularities in Continuous Space Word Representations
Mikolov, T., Yih, W.-T., & Zweig, G. (2013) · North American Chapter of the Association for Computational Linguistics · Cited by 2,912
Self-citationPhoneme recognition using time-delay neural networks
Waibel, A., Hanazawa, T., Hinton, G., Shikano, K., & Lang, K. J. (1989) · IEEE Transactions on Acoustics, Speech, and Signal Processing · doi:10.1109/29.21701 · Cited by 2,643 · cited 4×
“The LACE model thus differs from the earlier TDNN models [23], [55] in this attention masking, as well as the ResNet-like linear passthrough.”
BackgroundMethodInfluentialSWITCHBOARD: telephone speech corpus for research and development
Godfrey, J. J., Holliman, E. C., & McDaniel, J. (1992) · [Proceedings] ICASSP-92: 1992 IEEE International Conference on Acoustics, Speech, and Signal Processing · doi:10.1109/icassp.1992.225858 · Cited by 2,134 · cited 17×
“The 500-best lists from the lattices generated with the ResNet CNN system had an oracle (lowest achievable) WER of 2.7% on the Switchboard portion of the NIST 2000 evaluation set, and an oracle WER of 4.9% on the CallHome portion.”
BackgroundInfluentialDeep Speech 2: End-to-End Speech Recognition in English and Mandarin
Amodei, D., Anubhai, R., Battenberg, E., Case, C. T., Casper, J., Catanzaro, B., Chen, J., Chrzanowski, M., Coates, A., Diamos, G., et al. (2015) · arXiv (Cornell University) · doi:10.48550/arxiv.1512.02595 · Cited by 2,110
Learning deep structured semantic models for web search using clickthrough data
Huang, P.-S., He, X., Gao, J., Deng, L., Acero, A., & Heck, L. (2013) · Proceedings of the 22nd ACM international conference on Information & Knowledge Management · doi:10.1145/2505515.2505665 · Cited by 2,009
LSTM neural networks for language modeling
Sundermeyer, M., Schlüter, R., & Ney, H. (2012) · Interspeech 2012 · doi:10.21437/interspeech.2012-65 · Cited by 1,963 · cited 1×
“This is the case for both acoustic modeling [28], [29], [30], [31], [32], [33] and language modeling [34], [35], [36], [37].”
MethodCampbell, M., Hoane, A. J., & Hsu, F.-H. (2002) · Artificial Intelligence · doi:10.1016/s0004-3702(01)00129-1 · Cited by 1,226
The design for the wall street journal-based CSR corpus
Paul, D. B., & Baker, J. M. (1992) · Proceedings of the workshop on Speech and Natural Language - HLT '91 · doi:10.3115/1075527.1075614 · Cited by 1,099
Fiscus, J. G. (2002) · 1997 IEEE Workshop on Automatic Speech Recognition and Understanding Proceedings · doi:10.1109/asru.1997.659110 · Cited by 1,091 · cited 1×
“While ensemble classifier methods have long been popular in speech recognition systems [40]–”
MethodDeep convolutional neural networks for LVCSR
Sainath, T. N., Mohamed, A.-R., Kingsbury, B., & Ramabhadran, B. (2013) · 2013 IEEE International Conference on Acoustics, Speech and Signal Processing · doi:10.1109/icassp.2013.6639347 · Cited by 1,053 · cited 1×
“nition [79], [80], our networks are much deeper, and use linear bypass connections across convolutional layers.”
BackgroundPhonetic Feature Encoding in Human Superior Temporal Gyrus
Mesgarani, N., Cheung, C., Johnson, K., & Chang, E. F. (2014) · Science · doi:10.1126/science.1245994 · Cited by 993 · cited 1×
“Inspired by the human auditory cortex [57], where neighboring neurons tend to activate simultaneously, we employ a spatial smoothing technique to improve the accuracy of our”
MethodGeneralization of back-propagation to recurrent neural networks
Pineda, F. J. (1987) · Physical Review Letters · doi:10.1103/physrevlett.59.2229 · Cited by 931 · 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”
BackgroundSrivastava, R. K., Greff, K., & Schmidhuber, J. (2015) · arXiv · doi:10.48550/arxiv.1505.00387 · Cited by 912 · cited 1×
“, a linear transform of each layer’s input to the layer’s output [51], [52].”
BackgroundApplying Convolutional Neural Networks concepts to hybrid NN-HMM model for speech recognition
Abdel-Hamid, O., Mohamed, A.-R., Jiang, H., & Penn, G. (2012) · 2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2012.6288864 · Cited by 876 · cited 1×
“nition [79], [80], our networks are much deeper, and use linear bypass connections across convolutional layers.”
BackgroundConversational speech transcription using context-dependent deep neural networks
Seide, F., Li, G., & Yu, D. (2011) · Interspeech 2011 · doi:10.21437/interspeech.2011-169 · Cited by 852 · cited 7×
“To measure human performance, we leveraged an existing pipeline in which Microsoft data is transcribed on a weekly basis.”
BackgroundInfluentialSelf-citationPurely Sequence-Trained Neural Networks for ASR Based on Lattice-Free MMI
Povey, D., Peddinti, V., Galvez, D., Ghahremani, P., Manohar, V., Na, X., Wang, Y., & Khudanpur, S. (2016) · Interspeech 2016 · doi:10.21437/interspeech.2016-595 · Cited by 740 · cited 4×
“vious work [12], [64] by proposing the use of a mixed triphone/phoneme history in the language model.”
BackgroundMethodInfluentialSequence-discriminative training of deep neural networks
Veselý, K., Ghoshal, A., Burget, L., & Povey, D. (2013) · Interspeech 2013 · doi:10.21437/interspeech.2013-548 · Cited by 657 · cited 1×
“As noted in [62], [63], the necessary gradient for use in backpropagation is a simple function of the posterior probability of a particular acoustic model state at a given time, as computed by summing over all possible word sequences in an unconstrained manner.”
BackgroundUsing the Output Embedding to Improve Language Models
Press, O., & Wolf, L. (2017) · Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 2, Short Papers · doi:10.18653/v1/e17-2025 · Cited by 596
Speaker adaptation of neural network acoustic models using i-vectors
Saon, G., Soltau, H., Nahamoo, D., & Picheny, M. (2013) · 2013 IEEE Workshop on Automatic Speech Recognition and Understanding · doi:10.1109/asru.2013.6707705 · Cited by 577 · cited 1×
“tioning the network on an i-vector [59] characterization of each speaker [60], [61].”
BackgroundContext dependent recurrent neural network language model
Mikolov, T., & Zweig, G. (2012) · 2012 IEEE Spoken Language Technology Workshop (SLT) · doi:10.1109/slt.2012.6424228 · Cited by 560
Self-citationNeural Probabilistic Language Models
Bengio, Y., Schwenk, H., Senécal, J.-S., Morin, F., & Gauvain, J.-L. (2006) · Studies in Fuzziness and Soft Computing · doi:10.1007/10985687_6 · Cited by 540
Speech recognition by machines and humans
Lippmann, R. P. (1997) · Speech Communication · doi:10.1016/s0167-6393(97)00021-6 · Cited by 529 · cited 2×
“However, the error rate estimate in [19] is attributed to a “personal communication,” and the actual source of this number is ephemeral.”
BackgroundFrom Feedforward to Recurrent LSTM Neural Networks for Language Modeling
Sundermeyer, M., Ney, H., & Schluter, R. (2015) · IEEE/ACM Transactions on Audio, Speech, and Language Processing · doi:10.1109/taslp.2015.2400218 · Cited by 507 · cited 1×
“others [73], LSTMs outperform corresponding RNN-LMs.”
The Microsoft 2017 Conversational Speech Recognition System
Xiong, W., Wu, L., Alleva, F., Droppo, J., Huang, X., & Stolcke, A. (2018) · 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2018.8461870 · Cited by 472 · cited 1×
“which significantly boosts our bidirectional long shortterm memory (BLSTM) acoustic model performance 3) The use of long short-term memory (LSTM) rather than recurrent neural nets (RNNs) as language models, and the use of a letter-trigram input representation 4) A two-level system combination, based on a subsystem of…”
MethodSelf-citationThe Fisher Corpus: a Resource for the Next Generations of Speech-to-Text
Cieri, C., Miller, D. A. B., & Walker, K. P. (2004) · Language Resources and Evaluation Conference (LREC) · doi:10.63317/5q862povzthr · Cited by 464 · cited 8×
“This explains why we see many substitutions of backchannels for hesitations, but not vice-versa. of which comes from “quick transcription” [47] of the Fisher corpus, could also contribute to the frequency of this error type.”
BackgroundInfluentialSeide, F., Fu, H., Droppo, J., Li, G., & Yu, D. (2014) · Interspeech 2014 · doi:10.21437/interspeech.2014-274 · Cited by 443
Self-citationFast and accurate recurrent neural network acoustic models for speech recognition
Sak, H., Senior, A., Rao, K., & Beaufays, F. (2015) · Interspeech 2015 · doi:10.21437/interspeech.2015-350 · Cited by 385 · cited 1×
“We use a bidirectional architecture [56] without frame-skipping [29].”
MethodEnglish Conversational Telephone Speech Recognition by Humans and Machines
Saon, G., Kurata, G., Sercu, T., Audhkhasi, K., Thomas, S., Dimitriadis, D., Cui, X., Ramabhadran, B., Picheny, M., Lim, L.-L., et al. (2017) · Interspeech 2017 · doi:10.21437/interspeech.2017-405 · Cited by 349
Very Deep Convolutional Neural Networks for Noise Robust Speech Recognition
Qian, Y., Bi, M., Tan, T., & Yu, K. (2016) · IEEE/ACM Transactions on Audio, Speech, and Language Processing · doi:10.1109/taslp.2016.2602884 · Cited by 343 · cited 2×
“This is the case for both acoustic modeling [28]–[33] and language modeling [34]–[37].”
BackgroundResultThe DARPA 1000-word resource management database for continuous speech recognition
Price, P., Fisher, W. M., Bernstein, J., & Pallett, D. S. (2003) · ICASSP-88., International Conference on Acoustics, Speech, and Signal Processing · doi:10.1109/icassp.1988.196669 · Cited by 341 · cited 2×
“At first, the recognition tasks were simple ones like Resource Management [7] with a small vocabulary and carefully”
MethodAn Introduction to Computational Networks and the Computational Network Toolkit
Yu, D., Eversole, A., Seltzer, M., Yao, K., Kuchaiev, O., Zhang, Y., Seide, F. T. B., Huang, Z., Guenter, B. K., Wang, H., et al. (2014) · Cited by 332
Self-citationThe microsoft 2016 conversational speech recognition system
Xiong, W., Droppo, J., Huang, X., Seide, F., Seltzer, M., Stolcke, A., Yu, D., & Zweig, G. (2017) · 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2017.7953159 · Cited by 326
Self-citationPosterior probability decoding, confidence estimation and system combination
Evermann, G., & Woodland, P. (2000) · Cited by 233
Very deep multilingual convolutional neural networks for LVCSR
Sercu, T., Puhrsch, C., Kingsbury, B., & LeCun, Y. (2016) · 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2016.7472620 · Cited by 205 · cited 2×
“They are similar in spirit to those studied more recently by [31], [30], [61], [32], [33].”
MethodResultConsonant recognition by modular construction of large phonemic time-delay neural networks
Waibel, A., Sawai, H., & Shikano, K. (2003) · International Conference on Acoustics, Speech, and Signal Processing · doi:10.1109/icassp.1989.266376 · Cited by 153 · cited 1×
“TDNNmodels [23], [55] in this attentionmasking, as well as the ResNet-like linear pass-through.”
MethodTHE SRI MARCH 2000 HUB-5 CONVERSATIONAL SPEECH TRANSCRIPTION SYSTEM
Stolcke, A., Bratt, H., Butzberger, J., Franco, H., Gadde, V. R. R., Richey, C., Shriberg, E. E., Weng, F. Y., & Zheng, J. (2000) · Cited by 145
Self-citationA recurrent error propagation network speech recognition system
Robinson, T., & Fallside, F. (1991) · Computer Speech & Language · doi:10.1016/0885-2308(91)90010-n · Cited by 140 · 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”
BackgroundINSIGHTS INTO SPOKEN LANGUAGE GLEANED FROM PHONETIC TRANSCRIPTION OF THE SWITCHBOARD CORPUS
Greenberg, S., Hollenback, J., & Ellis, D. P. W. (1996) · Cited by 138
Advances in speech transcription at IBM under the DARPA EARS program
Chen, S. F., Kingsbury, B. E. D., Mangu, L., Povey, D., Saon, G., Soltau, H., & Zweig, G. (2006) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/tasl.2006.879814 · Cited by 125 · cited 5×
“The acoustic training data is comprised by LDC corpora 97S62, 2004S13, 2005S13, 2004S11 and 2004S09; see [12] for a full description.”
BackgroundMethodInfluentialSelf-citationThe IBM 2016 English Conversational Telephone Speech Recognition System
Saon, G., Sercu, T., Rennie, S., & Kuo, H.-K. J. (2016) · Interspeech 2016 · doi:10.21437/interspeech.2016-1460 · Cited by 96
The IBM 2015 English conversational telephone speech recognition system
Saon, G., Kuo, H.-K. J., Rennie, S., & Picheny, M. (2015) · Interspeech 2015 · doi:10.21437/interspeech.2015-632 · Cited by 91 · cited 1×
“They are similar in spirit to those studied more recently by [30]–[33], [61].”
ResultSuper-human multi-talker speech recognition: the IBM 2006 speech separation challenge system
Kristjansson, T. T., Hershey, J., Olsen, P. A., Rennie, S. J., & Gopinath, R. A. (2006) · Interspeech (USB) · doi:10.21437/interspeech.2006-25 · Cited by 87 · cited 1×
“carefully read newspaper speech [3] and rigidly constrained small-vocabulary tasks in noise [4], [5].”
BackgroundRecent innovations in speech-to-text transcription at SRI-ICSI-UW
Stolcke, A., Chen, B., Franco, H., Gadde, V. R. R., Graciarena, M., Hwang, M.-Y., Kirchhoff, K., Mandal, A., Morgan, N., Lei, X., et al. (2006) · IEEE Transactions on Audio, Speech and Language Processing · doi:10.1109/tasl.2006.879807 · Cited by 86 · cited 4×
“The rescored N-best lists from each subsystem are then aligned into a single confusion network [42] using the SRILM nbest-rover tool.”
BackgroundInfluentialSelf-citationDeep Convolutional Neural Networks with Layer-Wise Context Expansion and Attention
Yu, D., Xiong, W., Droppo, J., Stolcke, A., Ye, G., Li, J., & Zweig, G. (2016) · Interspeech 2016 · doi:10.21437/interspeech.2016-251 · Cited by 83 · cited 2×
“The last CNN variant is the LACE (layer-wise context expansion with attention) model [54], depicted in Fig.”
MethodSelf-citationA look at NIST'S benchmark ASR tests: past, present, and future
Pallett, D. S. (2004) · IEEE Workshop on Automatic Speech Recognition and Understanding (ASRU) · doi:10.1109/asru.2003.1318488 · Cited by 77 · cited 15×
“The 500-best lists from the lattices generated with the ResNet CNN system had an oracle (lowest achievable) WER of 2.7% on the Switchboard portion of the NIST 2000 evaluation set, and an oracle WER of 4.9% on the CallHome portion.”
MethodInfluentialChen, X., Liu, X., Qian, Y., Gales, M. J. F., & Woodland, P. C. (2016) · 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp.2016.7472829 · Cited by 77
Matsoukas, S., Gauvain, J.-L., Adda, G., Colthurst, T., Kao, C.-L., Kimball, O., Lamel, L. F., Lefèvre, F., Ma, J. Z., Makhoul, J., et al. (2006) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/tasl.2006.878257 · Cited by 55
Development of the 2003 CU-HTK conversational telephone speech transcription system
Evermann, G., Chan, H. Y., Gales, M. J. F., Hain, T., Liu, X., Mrva, D., Wang, L., & Woodland, P. C. (2004) · 2004 IEEE International Conference on Acoustics, Speech, and Signal Processing · doi:10.1109/icassp.2004.1325969 · Cited by 53 · cited 1×
“The history of work in this area includes key contributions by institutions such as IBM [12], BBN [13], SRI [14], AT&T [15], LIMSI [16], Cambridge University [17], Microsoft [18] and numerous others.”
BackgroundVery deep convolutional neural networks for LVCSR
Bi, M., Qian, Y., & Yu, K. (2015) · Interspeech (USB) · doi:10.21437/interspeech.2015-656 · Cited by 46 · cited 2×
“They are similar in spirit to those studied more recently by [31], [30], [61], [32], [33].”
MethodResultEfficient estimation of maximum entropy language models with n-gram features: an SRILM extension
Alumäe, T., & Kurimo, M. (2010) · Interspeech (USB) · doi:10.21437/interspeech.2010-519 · Cited by 43
Conversational telephone speech recognition
Gauvain, J.-L., Lamel, L. F., Schwenk, H., Adda, G., Chen, L., & Lefèvre, F. (2003) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2003.1198755 · Cited by 43 · cited 1×
“The history of work in this area includes key contributions by institutions such as IBM [12], BBN [13], SRI [14], AT&T [15], LIMSI [16], Cambridge University [17], Microsoft [18] and numerous others.”
BackgroundComparing Human and Machine Errors in Conversational Speech Transcription
Stolcke, A., & Droppo, J. (2017) · Interspeech 2017 · doi:10.21437/interspeech.2017-1544 · Cited by 41 · cited 2×
“which significantly boosts our bidirectional long shortterm memory (BLSTM) acoustic model performance 3) The use of long short-term memory (LSTM) rather than recurrent neural nets (RNNs) as language models, and the use of a letter-trigram input representation 4) A two-level system combination, based on a subsystem of…”
MethodSelf-citationStimulated Deep Neural Network for Speech Recognition
Wu, C., Karanasou, P., Gales, M., & Sim, K. C. (2016) · Interspeech (USB) · doi:10.21437/interspeech.2016-580 · Cited by 37 · cited 2×
“Stimulated networks [58] employ a similar approach.”
MethodSingle-channel mixed speech recognition using deep neural networks
Weng, C., Yu, D., Seltzer, M. L., & Droppo, J. (2014) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2014.6854681 · Cited by 33 · cited 1×
“carefully read newspaper speech [3] and rigidly constrained small-vocabulary tasks in noise [4], [5].”
BackgroundSelf-citationDeep Learning-Based Telephony Speech Recognition in the Wild
Han, K. J., Hahm, S., Kim, B., Kim, J., & Lane, I. R. (2017) · Interspeech (USB) · doi:10.21437/interspeech.2017-1695 · Cited by 25
Sequential classification criteria for NNs in automatic speech recognition
Wang, G., & Sim, K. C. (2011) · Interspeech (USB) · doi:10.21437/interspeech.2011-170 · Cited by 21 · cited 1×
“As noted in [62], [63], the necessary gradient for use in backpropagation is a simple function of the posterior probability of a particular acoustic model state at a given time, as computed by summing over all possible word sequences in an unconstrained manner.”
BackgroundImproving English Conversational Telephone Speech Recognition
Medennikov, I., Prudnikov, A., & Zatvornitskiy, A. (2016) · Interspeech (USB) · doi:10.21437/interspeech.2016-473 · Cited by 20 · cited 3×
“recognition system, considerably larger than previously reported for conversational speech recognition [37].”
BackgroundMethodLinearly augmented deep neural network
Ghahremani, P., Droppo, J., & Seltzer, M. L. (2016) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2016.7472646 · Cited by 19 · cited 1×
“, a linear transform of each layer’s input to the layer’s output [51], [52].”
BackgroundSelf-citationTranscription Methods for Consistency, Volume and Efficiency
Glenn, M. L., Strassel, S. M., Lee, H., Maeda, K., Zakhary, R., & Li, X. (2010) · Language Resources and Evaluation Conference (LREC) · doi:10.63317/3t2go3yc49ar · Cited by 16
Parallelizing WFST speech decoders
Mendis, C., Droppo, J., Maleki, S., Musuvathi, M., Mytkowicz, T., & Zweig, G. (2016) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2016.7472694 · Cited by 12 · cited 1×
“An initial decoding is done with a WFST decoder, using the architecture described in [65].”
MethodSelf-citationSelf-stabilized deep neural network
Ghahremani, P., & Droppo, J. (2016) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp.2016.7472719 · Cited by 8 · cited 1×
“) 5) Convergence was improved through a variation of selfstabilization [72], in which each output vector x of nonlinearities are scaled by (1)4 ln(1 + e 4β ), where a β is a scalar that is learned for each output.”
BackgroundSelf-citation(2014) · doi:10.4230/oasics.slate.2021.16 · Cited by 1 · 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.”
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
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.