Tony Robinson

• 1986: First implementation of Deep Neural Networks (Speech Recognition with Associative Networks.) - these had the form x_i = f(∑^i-1 w_ij x_j) and so were maximally deep yet trained well using all possible skip connections.
• 1987: First publication of Real Time Recurrent Learning (The utility driven dynamic error propagation network. also Static and dynamic error propagation networks with application to speech coding.)
• 1991: First state-of-the-art ASR with neural networks (Several improvements to a recurrent error propagation network phone recognition system)
• 1992: First real time large vocabulary continuous speech recognition system (Resource Management on DSP32C and SPARCstation)
• 1994: Shorten - the lossless audio compressor (SHORTEN: Simple lossless and near-lossless waveform compression. and Shorten_(file_format))
• 1996: First end-to-end training of neural nets and HMMs Forward-backward retraining of recurrent neural networks
• 1999: The time-first decoder (Recognition of sequential data using finite state sequence models organized in a tree structure. and Time-first search for large vocabulary speech recognition.)
• As supervisor to MPhil students:
  • First speech editor - edit audio as text
  • First editor for correcting speech recognition transcripts
  • First automatically scrolling teleprompter (Autocue)

• Speech processing system and method. T. W. J. Ash and A. J. Robinson. Patent application PCT/GB2016/053456. November 2016.
• Scaling Recurrent Neural Network Language Models. Will Williams, Niranjani Prasad, David Mrva, Tom Ash and Tony Robinson. In Proc. ICASSP, pages 5391-5395, 2015.
• Audio coding systems and methods. R. C. F. Tucker, C. W. Saymour and A. J. Robinson. Patent US6675144. January 2004.
• Connectionist speech recognition of broadcast news. A. J. Robinson, G. D. Cook, D. P. W. Ellis, E. Fosler-Lussier, S. J. Renals, and D. A. G. Williams. Speech Communication, 37(1), 2002.
• Adaptive model-based speech enhancement. Beth Logan and Tony Robinson. Speech Communication, 34(4), July 2001.
• Improved language modelling though better language model evaluation measures. Philip Clarkson and Tony Robinson. Computer Speech and Language, 15(1), January 2001.
• Indexing and retrieval of broadcast news. Steve Renals, Dave Abberley, David Kirby and Tony Robinson. Speech Communication, 32(1):5-20, 2000.
• Segmentation of a speech waveform according to glottal open and closed phases using an autoregressive-HMM. Gavin Smith and Tony Robinson. In Proceedings of the International Conference on Spoken Language Processing, 2000.
• Subspace techniques in speech enhancement. Gavin Smith, Mahesan Niranjan and Tony Robinson. In Neural Networks in Signal Processing 9, 1999.
• Recognition of sequential data using finite state sequence models organized in a tree structure. A. J. Robinson. Patent US5983180. November 1999.
• Modelling the singing voice. Gavin Smith, Tony Robinson, and Mahesan Niranjan. In 1999 Cambridge Music Processing Colloquium, 1999.
• Towards improved language model evaluation measures. Philip Clarkson and Tony Robinson. In Proceedings of the European Conference on Speech Technology, volume 5, pages 1927-1930, September 1999.
• Speech coding using mixture of Gaussians polynomial model. Parham Zolfaghari and Tony Robinson. In Proceedings of the European Conference on Speech Technology, volume 3, pages 1495-1498, September 1999.
• Recognition, indexing and retrieval of british broadcast news with the THISL system. Tony Robinson, Dave Abberley, David Kirby, and Steve Renals. In Proceedings of the European Conference on Speech Technology, volume 3, pages 1267-1270, September 1999.
• Compression of acoustic features - are perceptual quality and recognition performance incompatible goals?. Roger Tucker, Tony Robinson, James Christie and Carl Seymour. In Proceedings of the European Conference on Speech Technology, volume 5, pages 2155-2158, September 1999.
• Accessing Information in Spoken Audio, ESCA Tutorial and Research Workshop (ITRW). Tony Robinson and Steve Renals (eds.), Cambridge, UK, April 18-19, 1999.
• The THISL broadcast news retrieval system. Dave Abberley, David Kirby, Steve Renals and Tony Robinson. In Proceedings of the ESCA workshop: Accessing information in spoken audio, pages 14-19. Cambridge University, April 1999.
• Recognition-compatible speech compression for stored speech. Roger Tucker, Tony Robinson, James Christie, and Carl Seymour. In Proceedings of the ESCA workshop: Accessing information in spoken audio, pages 69-72. Cambridge University, April 1999.
• Summarisation of spoken audio through information extraction. Robin Valenza, Tony Robinson, Marianne Hickey and Roger Tucker. In Proceedings of the ESCA workshop: Accessing information in spoken audio, pages 111-116. Cambridge University, April 1999.
• An overview of the Sprach system for the transcription of broadcast news. Gary Cook, James Christie, Dan Ellis, Eric Fosler-Lussier, Yoshi Gotoh, Brian Kingsbury, Nelson Morgan, Steve Renals, Tony Robinson, and Gethin Williams. In Proceedings of the DARPA Broadcast News Workshop. Morgan Kaufmann, 1999.
• Retrieval of broadcast news documents with the THISL system. Dave Abberley, Steve Renals, Gary Cook and Tony Robinson. In Proc. Seventh Text Retrieval Conference (TREC-7), 1999.
• G. A. Smith, A. J. Robinson, and M. Niranjan. Speech modelling: Models, parameter estimation and its application to speech enhancement. Technical Report CUED/F-INFENG/TR.345, Cambridge University Engineering Dept, Cambridge, UK, 1999.
• A practical perceptual frequency autoregressive HMM enhancement system. Beth Logan and Tony Robinson. In Proceedings of the International Conference on Spoken Language Processing, November 1998.
• The applicability of adaptive language models to the broadcast news task. Philip Clarkson and Tony Robinson. In Proceedings of the International Conference on Spoken Language Processing, November 1998.
• Real-time recognition of broadcast news. Gary Cook, Tony Robinson, and James Christie. In Proceedings of the International Conference on Spoken Language Processing, November 1998.
• G. D. Cook, A. J. Robinson, and J. deM. Christie. Connectionist acoustic modelling in the Abbot system. In Institute of Acoustics Autumn conference on Speech and Hearing, November 1998.
• Andrew Senior and Tony Robinson. Online cursive handwriting recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 20(3):309-321, 1998.
• Time-first search for large vocabulary speech recognition. Tony Robinson and James Christie. In Proc. ICASSP, pages 829-832, 1998.
• Gary Cook and Tony Robinson. Transcribing broadcast news with the 1997 Abbot system. In Proc. ICASSP, pages 917-920, 1998.
• Joint Prediction and Vector Quantisation. Carl Seymour and Tony Robinson. 1997.
• The 1997 Abbot system for the transcription of broadcast news. G. D. Cook and A. J. Robinson. In Proc. of the Broadcast News Transcription and Understanding workshop, pages 49-54. Morgan Kaufmann, February 1998.
• Dave Abberley, Steve Renals, Gary Cook, and Tony Robinson. The THISL spoken document retrieval system. In TREC-6 Proceedings , 1998.
• Low bit rate audio coder and decoder operating in a transform domain using vector quantization. A. J. Robinson. Patent US5999899. October 1997.
• S. J. Young, M. Adda-Dekker, X. Aubert, C. Dugast, J.-L. Gauvain, D. J. Kershaw, L. Lamel, D. A. Leeuwen, D. Pye, A. J. Robinson, H. J. M. Steeneken, and P. C. Woodland. Multilingual large vocabulary speech recognition: the European sqale project. Computer Speech and Language , 11:73-89, 1997.
• Transcription of broadcast television and radio news: The 1996 Abbot system. G. D. Cook, D. J. Kershaw, J. D. M. Christie, and A. J. Robinson. In Proc. of DARPA Speech Recognition Workshop , pages 79-84. Morgan Kaufmann, February 1997.
• Ensemble methods for connectionist acoustic modelling. G. D. Cook, S. R. Waterhouse, and A. J. Robinson. In Proceedings of the European Conference on Speech Technology , volume 4, pages 1959-1962, September 1997.
• C. W. Seymour and A. J. Robinson. A low-bit-rate speech coder using adaptive line spectral frequency prediction. In Proceedings of the European Conference on Speech Technology , volume 3, pages 1319-1322, September 1997.
• A segmental formant vocoder based on linearly varying mixtures of Gaussians. Parham Zolfaghari and Tony Robinson. In Proceedings of the European Conference on Speech Technology, volume 1, pages 425-428, September 1997.
• Improving autoregressive hidden Markov model recognition accuracy using a non-linear frequency scale with application to speech enhancement. B. T. Logan and A. J. Robinson. In Proceedings of the European Conference on Speech Technology, volume 4, pages 2103-2106, September 1997.
• Transcription of broadcast television and radio news: The 1996 Abbot system. G. D. Cook, D. J. Kershaw, J. D. M. Christie, C. W. Seymour and S. R. Waterhouse. In Proc. ICASSP, 1997.
• Language model adaptation using mixtures and an exponentially decaying cache. Philip Clarkson and Tony Robinson. In Proc. ICASSP, pages 799-802, 1997.
• Enhancement and recognition of noisy speech within an autoregressive hidden Markov model framework using noise estimates from the noisy signal. B. T. Logan and A. J. Robinson. In Proc. ICASSP, 1997.
• A formant vocoder based on mixtures of Gaussians. Parham Zolfaghari and Tony Robinson. In Proc. ICASSP, volume 2, pages 1575-1578, 1997.
• Noise estimation for enhancement and recognition within an autoregressive hidden Markov model framework. B. T. Logan and A. J. Robinson. In The Sixth Australian International Conference on Speech Science and Technology , pages 85-90, December 1996.
• The 1995 ABBOT LVCSR system for multiple unknown microphones. Dan Kershaw, Tony Robinson, and Steve Renals. In Proceedings of the International Conference on Spoken Language Processing, volume 3, pages 1325-1328, October 1996.
• Smoothed local adaptation of connectionist systems. Steve Waterhouse, Dan Kershaw, and Tony Robinson. In Proceedings of the International Conference on Spoken Language Processing, volume 3, pages 1313-1316, October 1996.
• Boosting the performance of connectionist large vocabulary speech recognition. Gary Cook and Tony Robinson. In Proceedings of the International Conference on Spoken Language Processing, pages 1305-1308, October 1996.
• Formant analysis using mixtures of Gaussians.Parham Zolfaghari and Tony Robinson. In Proceedings of the International Conference on Spoken Language Processing, volume 2, pages 1229-1232, October 1996.
• Bayesian methods for mixtures of experts. Steve Waterhouse, David MacKay and Tony Robinson. In D. S. Touretzky, M. C. Mozer, and M. E. Hasselmo, editors, Advances in Neural Information Processing Systems 8. Morgan Kaufmann, 1996.
• Constructive algorithms for hierarchical mixtures of experts. S. R. Waterhouse and A. J. Robinson. In D. S. Touretzky, M. C. Mozer, and M. E. Hasselmo, editors, Advances in Neural Information Processing Systems 8. Morgan Kaufmann, 1996.
• Context-dependent classes in a hybrid recurrent network-HMM speech recognition system. Dan Kershaw, Tony Robinson, and Mike Hochberg. In D. S. Touretzky, M. C. Mozer, and M. E. Hasselmo, editors, Advances in Neural Information Processing Systems 8, pages 750-756. Morgan Kaufmann, 1996.
• Forward-backward retraining of recurrent neural networks. Andrew Senior and Tony Robinson. In D. S. Touretzky, M. C. Mozer and M. E. Hasselmo, editors, Advances in Neural Information Processing Systems 8. Morgan Kaufmann, 1996.
• Real-time recognition of broadcast radio speech. G. D. Cook, J. D. Christie, P. R. Clarkson, M. M. Hochberg, B. T. Logan, A. J. Robinson, and C. W. Seymour. In Proc. ICASSP, 1996.
• The 1995 ABBOT hybrid connectionist-HMM large vocabulary recognition system. D. J. Kershaw, A. J. Robinson, and S. J. Renals. In Speech Recognition Workshop. Morgan Kaufmann, February 1996. ISBN 1-55860-422-7.
• The use of recurrent networks in continuous speech recognition. Tony Robinson, Mike Hochberg, and Steve Renals. In Chin-Hui Lee, Frank K. Soong, and Kuldip K. Paliwal, editors, Automatic Speech and Speaker Recognition: Advanced Topics, Chapter 10. Kluwer Academic Publishers, 1996.
• Context-dependent modelling in the ABBOT LVSCR system. D. J. Kershaw, M. M. Hochberg, and A. J. Robinson. In Proceedings IEEE ASR workshop at Snowbird, December 1995.
• D. J. Kershaw, M. M. Hochberg, and A. J. Robinson. Context-dependent classes in a hybrid recurrent network-HMM speech recognition system. Technical Report CUED/F-INFENG/TR.217, Cambridge University Engineering Department, June 1995.
• Utterance clustering for large vocabulary speech recognition. G. D. Cook and A. J. Robinson. In Proceedings of the European Conference on Speech Technology, September 1995.
• Speaker-adaptation for hybrid HMM-ANN continuous speech recognition system. João Neto, Luis Almeida, Mike Hochberg, Ciro Martins, Luis Nunes, Steve Renals, and Tony Robinson. In Proceedings of the European Conference on Speech Technology, 1995.
• Steven R. Waterhouse and A. J. Robinson. Pruning and growing hierarchical mixtures of experts. In Proceedings of the IEE conference on Artificial Neural Networks , 1995.
• Gary Cook and Tony Robinson. Training MLPs via the expectation-maximisation algorithm. In Proceedings of the IEE conference on Artificial Neural Networks , 1995.
• WSJCAM0: A British English speech corpus for large vocabulary continuous speech recognition. Tony Robinson, Jeroen Fransen, David Pye, Jonathan Foote, and Steve Renals. In Proc. ICASSP, pages 81-84, 1995.
• The 1994 Abbot hybrid connectionist/HMM large-vocabulary recognition system. M. M. Hochberg, G. D. Cook, S. J. Renals, A. J. Robinson, and R. T. Schechtman. In ARPA Spoken Language Systems, pages 170-175, 1995.
• M. M. Hochberg, S. J. Renals, A. J. Robinson, and G. D. Cook. Recent improvements to the Abbot large vocabulary CSR system. In Proc. ICASSP , pages 401-404, 1995.
• Non-linear prediction of acoustic vectors using hierarchical mixtures of experts. S. R. Waterhouse and A. J. Robinson. In G. Tesauro J. Cowan and J. Alspector, editors, Advances in Neural Information Processing Systems 7. Morgan Kaufmann, 1995.
• E. M. Ellis and A. J. Robinson. Spiral: A vibrotactile based speech listening aid. In Third International Conference on Tactile Aids, Hearing Aids and Cochlear Implants , Miami, USA, May 1994.
• M. M. Hochberg, S. J. Renals, and A. J. Robinson. ABBOT: the CUED hybrid conectionist-HMM large vocabulary recognition system. In Proc. of SLS Workshop , Princeton, March 1994. ARPA, Morgan Kaufmann.
• SHORTEN: Simple lossless and near-lossless waveform compression. Tony Robinson. Technical Report CUED/F-INFENG/TR.156, Cambridge University Engineering Department, 1994.
• Connectionist model combination for large vocabulary speech recognition. Mike Hochberg, Gary Cook, Steve Renals, and Tony Robinson. In IEEE Workshop on Neural Networks for Signal Processing, pages 269-278, 1994.
• Mike Hochberg, Tony Robinson, Steve Renals, and Dan Kershaw. Large vocabulary continuous speech recognition using a hybrid connectionist-HMM system. In Proceedings of the International Conference on Spoken Language Processing, pages 1499-1502, 1994.
• Classification using hierarhical mixtures of experts. S. R. Waterhouse and A. J. Robinson. In IEEE Workshop on Neural Networks for Signal Processing, pages 177-186, 1994.
• WSJCAM0 corpus and recording description. Jeroen Fransen, Dave Pye, Tony Robinson, Phil Woodland, and Steve Young. Technical Report CUED/F-INFENG/TR.192, Cambridge University Engineering Department, September 1994.
• John Garofolo, Tony Robinson, and Jonathan Fiscus. The development of file formats for very large speech corpora: Sphere and shorten. In Proc. ICASSP , volume I, pages 113-116, 1994.
• Tony Robinson, Mike Hochberg, and Steve Renals. IPA: Improved phone modelling with recurrent neural networks. In Proc. ICASSP, 1994.
• The application of recurrent nets to phone probability estimation. Tony Robinson. IEEE Transactions on Neural Networks, 5(2), March 1994.
• Learning temporal dependencies in connectionist speech recognition. Steve Renals, Mike Hochberg and Tony Robinson. In J. D. Cowan, G. Tesauro and J. Alspector, editors, Advances in Neural Information Processing Systems, volume 6, pages 1051-1058. Morgan Kaufmann, San Mateo CA, 1994.
• A neural network based, speaker independent, large vocabulary, continuous speech recognition system: The WERNICKE project. Tony Robinson, Luis Almeida, Jean-Marc Boite, Herve Bourlard, Frank Fallside, Mike Hochberg, Dan Kershaw, Phil Kohn, Yochai Konig, Steve Renals, Marco Saerens, João Paulo Neto, Nelson Morgan, and Chuck Wooters. In Proceedings of the European Conference on Speech Technology, 1993.
• Speech synthesis using artificial neural networks trained on cepstral coefficients. Christine Tuerk and Tony Robinson. In Proceedings of the European Conference on Speech Technology, September 1993.
• A new frequency shift function for reducing inter-speaker variance. Christine Tuerk and Tony Robinson. In Proceedings of the European Conference on Speech Technology, September 1993.
• E. M. Ellis and A. J. Robinson. A tactile system for speech listening based on phonetic representation. In ESCA Conference on Speech and Language Technology for Disabled Persons, Stockholm, 31 May - 2 June 1993.
• A phonetic tactile speech listening listening system. E. M. Ellis and A. J. Robinson. Technical Report CUED/F-INFENG/TR.122, Cambridge University Engineering Department, May 1993.
• C. Giguère, P. C. Woodland, and A. J Robinson. Application of an auditory model to the computer simulation of hearing impairment: Preliminary results. Canadian Acoustics, 21(3), 1993.
• Tony Robinson. Artifcial Neural Networks: The mole-grips of the speech scientist. In Visual Representations of Speech Signals. John Wiley and Sons, 1993.
• Tony Robinson. The state space and “ideal input” representations of recurrent networks. In Visual Representations of Speech Signals , pages 327-334. John Wiley and Sons, 1993.
• Tony Robinson. Recurrent nets for phone probability estimation. In Proceedings of the ARPA Continuous Speech Recognition Workshop , Stanford, September 1992.
• Tony Robinson. Practical network design and implementation. In Proceedings of the Cambridge Neural Network Summer School , Cambridge Programme for Industry, Cambridge University, September 1992.
• A real-time recurrent error propagation network word recognition system. Tony Robinson. In Proc. ICASSP, volume I, pages 617-620, 1992.
• Anne Cutler and Tony Robinson. Response time as a metric for comparison of speech recognition by humans and machines. In Proceedings of the International Conference on Spoken Language Processing, October 1992.
• Christine Tuerk and Tony Robinson. A multiple-speaker phoneme durational model. In Institute of Acoustics Autumn Conference on Speech and Hearing , November 1992.
• Errol M. Ellis and Tony Robinson. Two dimensional representation of phonemes of the English language. In Institute of Acoustics Autumn Conference on Speech and Hearing , November 1992.
• Several improvements to a recurrent error propagation network phone recognition system. Tony Robinson. Technical Report CUED/F-INFENG/TR.82, Cambridge University Engineering Department, September 1991.
• Christine Tuerk, Peter Monaco, and Tony Robinson. The development of a connectionist multiple-voice text-to-speech system. In Proc. ICASSP , 1991.
• N. H. Russell, F. Fallside, A. J. Robinson, and R. W. Prager. Lexical access using a recurrent error propagation network. In Proceedings of the European Conference on Speech Technology, Genoa, Italy, 1991.
• Tony Robinson. Recognition of continuous speech using recurrent error propagation networks. In Proceedings of Voice Systems Worldwide , London, June 1991.
• A recurrent error propagation network speech recognition system. Tony Robinson and Frank Fallside. Computer Speech and Language , 5(3):259-274, July 1991.
• Tony Robinson and Frank Fallside. Word recognition from the DARPA resource management database with the Cambridge recurrent error propagation network speech recognition system. In Third Australian International Conference on Speech Science and Technology, Melbourne, November 1990.
• A comparison of preprocessors for the Cambridge recurrent error propagation network speech recognition system. Tony Robinson, John Holdsworth, Roy Patterson, and Frank Fallside. In Proceedings of the International Conference on Spoken Language Processing, pages 1033-1036, Kobe, Japan, November 1990.
• F. Fallside, H. Lucke, T. P. Marsland, P. J. O'Shea, M. St. J. Owen, R. W. Prager, A. J. Robinson, and N. H. Russell. Continuous speech recognition for the TIMIT database using neural networks. In Proc. ICASSP, 1990.
• Phoneme recognition from the TIMIT database using recurrent error propagation networks. Tony Robinson and Frank Fallside. Technical Report CUED/F-INFENG/TR.42, Cambridge University Engineering Department, March 1990.
• Tony Robinson and Frank Fallside. Dynamic reinforcement driven error propagation networks with application to game playing. In Eleventh Annual Conference of the Cognitive Science Society, pages 836-843, 1989.
• Tony Robinson. Experiments in the non-linear predictive coding of speech. Internal report, NTT Basic Research Laboratories, 1989.
• Dynamic Error Propagation Networks. A. J. Robinson. PhD thesis, Cambridge University Engineering Department, February 1989.
• A. J. Robinson and F. Fallside. A dynamic connectionist model for phoneme recognition. In Neural Networks from Models to Applications: Proceedings of nEuro'88 , pages 541-550. I.D.S.E.T., Paris, 1989.
• A. J. Robinson, M. Niranjan, and F. Fallside. Generalising the nodes of the error propagation network. Technical Report CUED/F-INFENG/TR.25, Cambridge University Engineering Department, 1988.
• Static and dynamic error propagation networks with application to speech coding. A. J. Robinson and F. Fallside. In Dana Z. Anderson, editor, Neural Information Processing Systems. American Institute of Physics, 1988.
• F. Fallside, T. D. Harrison, R. W. Prager, and A. J. Robinson. A comparison of three connectionist models for phoneme recognition in continuous speech. In ATR Workshop on Neural Networks and Parallel Distributed Processing , Osaka, Japan, July 1988. ATR Auditory and Visual Perception Research Laboratories and ATR Interpreting Telephony Research Laboratories.
• A. J. Robinson and F. Fallside. A dynamic connectionist model for phoneme recognition: Preliminary results. Technical Report CUED/F-INFENG/TR.14, Cambridge University Engineering Department, 1988.
• The utility driven dynamic error propagation network. A. J. Robinson and F. Fallside. Technical Report CUED/F-INFENG/TR.1, Cambridge University Engineering Department, 1987.
• Speech Recognition with Associative Networks. Tony Robinson, MPhil thesis, Cambridge University Engineering Department, Trumpington Street, Cambridge CB2 1PZ, UK, September 1986.

IN PROGRESS: Email me if you want a copy of something that doesn't yet have a PDF. See also Google Scholar, ResearchGate and Semantic Scholar.