Professor Leslie Smith's Papers

...and some posters and presentations too

If you have any trouble downloading any of these, please email me (address at bottom of page), and I will send you a .pdf.

Entropy, published 4 December 2020. DOI 10.3390/e22121367
M.Pahar, L.S. Smith Coding and Decoding Speech using a Biologically Inspired Coding System: presented at IEEE SSCI 2020, (virtual conference) 1-4 December 2020. DOI 10.1109/SSCI47803.2020.9308328.
Leslie S. Smith talk at Brain Informatics 2020: .pdf of slides, actual presentation (52 minutes long): Invited talk at the Brain Informatics 2020 Conference, 19 September 2020
Leslie S. Smith Comments on Sejnowski's "The unreasonable effectiveness of deep learning in artificial intelligence": ArXiv preprint [arXiv:2002.04806], submitted Fri, 20 Mar 2020
Graham S. Wood, Alberto Torin, Asaad K. Al-mashaal, Leslie S. Smith, Enrico Mastropaolo, Michael J. Newton, and Rebecca Cheung Design and Characterization of a Micro-Fabricated Graphene-Based MEMS Microphone: IEEE Sensors Journal, 19(7), September 2019, 7234 - 7242, DOI 10.1109/JSEN.2019.2914401. It is with huge sadness that I note the death of Enrico Mastropaolo, at a very young age.
Andrew Abel, Chengxiang Gao, Leslie Smith, Roger Watt, Amir Hussain Fast Lip Feature Extraction Using Psychologically Motivated Gabor Features: 2018 IEEE Symposium Series on Computational Intelligence (SSCI), 18-21 November 2018, Bangalore, India, DOI 10.1109/SSCI.2018.8628931, Added to IEEE Xplore: 31 January 2019. Non-final version.
L.S. Smith Deep neural networks: the only show in town?: A position paper for the workshop Can Deep Neural Networks (DNNs) provide the basis for Artificial General Intelligence (AGI) at the Artificial General Intelligence 2016 conference, July 15-18, 2016, New York, USA.
Areli Rojo-Hernandeza, Giovanny Sanchez-Rivera, Gerardo Avalos-Ochoa, Hector Perez-Meana, Leslie S. Smith A Compact Digital Gamma-tone Filter Processor (also available here): Microprocessors and Microsystems, 45A , 216-225 August 2016: doi:10.1016/j.micpro.2016.05.010
Varley A, Tyler A, Smith L, Dale P & Davies M, Mapping the spatial distribution and activity of 226Ra at legacy sites through Machine Learning interpretation of gamma-ray spectrometry data,: Science of the Total Environment, 545-546, pp. 654-661, 2016, 10.1016/j.scitotenv.2015.10.112
L.S. Smith Towards a Neuromorphic Microphone: Frontiers in Neuroscience (Neuromorphic Section), 9:398, 26 October 2015, doi 10.3389/fnins.2015.00398.
Victoria Hodge, Mark Jessop, Michael Weeks, Aaron Turner, Tom Jackson, Colin Ingram, Leslie Smith, Jim Austin, A Digital Repository and Execution Platform for Interactive Scholarly Publications in Neuroscience: Neuroinformatics, published online 26 August 2015, 10.1007/s12021-015-9276-3. 18 pages.
Leslie S. Smith, Why sharing matters for electrophysiological data analysis: Brain Research Bulletin, published online 3 July 2015, Volume 119, October 2015, Part B, 145-149, doi:10.1016/j.brainresbull.2015.06.009
Adam Varley, Andrew Tyler, Leslie Smith, Paul Dale, Mike Davies: Remediating radium contaminated legacy sites: Advances made through machine learning in routine monitoring of "hot" particles: Science of the Total Environment, 521-522,(2015), 271-279
A. Abel, D. Hunter, L.S.Smith, A biologically inspired onset and offset speech segmentation approach: presented at IJCNN 2015, Killarney, Ireland, 12-17 July 2015. doi: 10.1109/IJCNN.2015.7280347. Watch Dr Andrew Abel talking about this work at DemoFest 2014 in Edinburgh.
Adam Varley, Andrew Tyler, Leslie Smith, Paul Dale, Development of a neural network approach to characterise ²²⁶Ra contamination at legacy sites using gamma-ray spectra taken from boreholes: Journal of Environmental Radioactivity, 140 (2015) 130-140
K. Swingler and L.S. Smith, An Analysis of the Local Optima Storage Capacity of Hopfield Network Based Fitness Function Models: Transactions on Computational Collective Intelligence XVII, N.T. Nguyen, R. Kowalcyk, A. Fred, F. Joaquim (eds), Springer 2014, pp 248-271
Abdulrahman Alalshekmubarak and Leslie S. Smith On Improving the Classification Capability of Reservoir Computing for Arabic Speech Recognition: in Wermter, S., Weber, C., Duch, W., Honkela, T., Koprinkova-Hristova, P., Magg, S., Palm, G., Villa, A.E.P. (Eds.) , Artificial Neural Networks and Machine Learning-ICANN 2014, 24th International Conference on Artificial Neural Networks, Lecture Notes in Computer Science 8681, Springer Heidelberg, 2014, pages 225-232. Here is the poster as presented.
Leslie S. Smith, Jim Austin, Stephen Eglen, Tom Jackson, Mark Jessop, Bojian Liang, Michael Weeks and Evelyne Sernagor, The CARMEN data sharing portal project: what have we learned? There's also a presentation given at the Research Data Association meeting in Paris in September 2015.: Neuroinformatics 2014, Leiden, The Netherlands, 25-27 August 2014. Here is the poster as presented.
S. Wang, T.J. Koickal, A. Hamilton, R. Cheung, L.S. Smith, A Bio-Realistic Analog CMOS Cochlea Filter With High Tunability and Ultra-Steep Roll-Off: IEEE Transactions on Biomedical Circuits and Systems, 9(3), 297-311, 2015 ( e-published 31 July 2014, doi 10.1109/TBCAS.2014.2328321).
A. Alalshekmubarak, L.S. Smith, A noise robust Arabic speech recognition system based on the echo state network: Acoustical Society of America 167th meeting, Providence RI, USA, 5-9 May 2014. (J. Acoustical Society of America, 135 (4) part 2, p2195) Poster .pdf.
L.S. Smith, A. Abel, Spectrotemporal Gabor filters for feature detection: Acoustical Society of America 167th meeting, Providence RI, USA, 5-9 May 2014. (J. Acoustical Society of America, 135 (4) part 2, p2297) Slides .pdf
K. Swingler, L.S. Smith, Training and making calculations with mixed order hyper-networks: Neurocomputing, 2014. Available online, 8 April 2014. doi http://dx.doi.org/10.1016/j.neucom.2013.11.041
Kevin Swingler, Leslie S. Smith Mixed Order Associative Networks for Function Approximation, Optimisation and Sampling: In: ESANN 2013 proceedings, 21st European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN. 21st European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2013, 24.4.2013 - 26.4.2013, Bruges, Belgium, pp. 23-28.
Alalshekmubarak, Abdulrahman; Smith, Leslie S, A novel approach combining recurrent neural network and support vector machines for time series classification. (preprint): Innovations in Information Technology (IIT), 2013 9th International Conference on, pp.42,47, 17-19 March 2013 doi: 10.1109/Innovations.2013.6544391
L.S. Smith, Perceptual time, perceptual reality and general intelligence.: in Artificial General Intelligence, Proc of 5th International Conference, Oxford, December 2012, J. Bach, B. Goertzel and M. Ikle (editors), LNAI 7716, Springer, pp 292-301, 2012.
L.S. Smith, On the relationship between neural coding and the perception of the present moment.: poster presented at the Society for Neuroscience Annual Meeting, New Orleans, October 2012
Crook SM, Bednar JA, Berger S, Cannon R, Davison AP, Djurfeldt M, Eppler J, Kriener B, Furber S, Graham B, Plesser HE, Schwabe L, Smith L, Steuber V, van Albada S., Creating, documenting and sharing network models. (pre-publication pdf is available): Network: Computation in Neural Systems, 2012, 1-19, (ePub) DOI: 10.3109/0954898X.2012.722743
Plamen L. Simeonov, Edwin H. Brezina, Ron Cottam, Andree C. Ehresmann, Arran Gare, Ted Goranson, Jaime Gomez-Ramirez, Brian D. Josephson, Bruno Marchal, Koichiro Matsuno, Robert S. Root-Bernstein, Otto E. Rossler, Stanley N. Salthe, Marcin Schroeder, Bill Seaman, Pridi Siregar, Leslie S. Smith: Stepping Beyond the Newtonian Paradigm in Biology: Towards an Integrable Model of Life: Accelerating Discovery in the Biological Foundations of Science (INBIOSA White Paper),: in Simeonov P.L., Smith L.S., Ehresmann, A.C. (eds) : Integral Biomathics: Tracing the road to reality, Springer Verlag, pp319-418, 2012.
Simeonov P.L., Smith L.S., Ehresmann, A.C. (eds), Integral Biomathics: Tracing the road to reality: Springer Verlag, 2012. Note that I have spare copies of this book (as at December 2012): email me if you would like one.
Friedrich Sommer, Thomas Wachtler, Andrew Davison, Michael Denker, Jeffrey Grethe, Sonja Grün, Kenneth Harris, Colin Ingram, Marja-Leena Linne, Bengt Ljungquist, John Miller, Roman Moucek, Hyrum Sessions, Gordon Shepherd, Leslie Smith, Jeff Teeters and Shiro Usui, Mission and activities of the INCF Electrophysiology Data Sharing Task Force (abstract): Front. Neuroinform. Conference Abstract: 5th INCF Congress of Neuroinformatics, 10-12 September 2012, Munich. doi: 10.3389/conf.fninf.2014.08.00088
M. Newton, L.S. Smith A neurally-inspired musical instrument classification system based upon the sound onset (note: paper is not final version): Journal of the Acoustical Society of America, Volume 131, Issue 6, pp. 4785-4798 June 2012: doi 10.1121/1.4707535
Bo Yu, Terrence Mak, Xiangyu Li, Leslie Smith, Yihe Sun and Chi-Sang Poon, Stream-based Hebbian Eigenfilter for real-time neuronal spike discrimination: BioMedical Engineering OnLine 2012, 11:18 doi:10.1186/1475-925X-11-18 Published: 10 April 2012
L.S. Smith: Seminar given at Imperial College, London, on December 7 2011, entitled "Am I Spiking neurons?": Powerpoint presentation slides (pdf)
R. Latif, E. Mastropaulo, A. Bunting, R. Cheung, T. Koickal, A. Hamilton, M. Newton, L. Smith, Low frequency tantalum electromechanical systems for biomimetical applications: J. Vac. Sci. Technol. B 29(6), Nov/Dec 2011. DOI 10.1116/1.3662408
Bo Yu, Terrence Mak, Alex Yakovlev, Chi-Sang Poon, Yihe Sun, Leslie Samuel Smith: Memory Efficient On-Line Streaming for Multichannel Spike Train Analysis: 33rd Annual International IEEE EMBS Conference, Aug 30-Sept 1, 2011, Boston, pp 2315-2318,
M. J. Newton and L.S. Smith, Biologically-inspired neural coding of sound onset for a musical sound classification task,: Neural Networks (IJCNN), The 2011 International Joint Conference on, 1386-1393, 10.1109/IJCNN.2011.6033386.
Iffat Gheyas, Leslie Smith, A novel neural network ensemble architecture for time series forecasting: Neurocomputing 74, 3855-3864 (2011), doi:10.1016/j.neucom.2011.08.005.
Leslie Smith, Daniel Metz, Jungpen Bao, and Pedro Bizarro, Events, Neural Systems and Time Series: in M. Cezon and Y. Wolfsthal (Eds.): ServiceWave 2011 Workshops, LNCS 6569, pp. 196--202. Springer, Heidelberg (2011).
Plamen L. Simeonov, Andree C. Ehresmann, Leslie S. Smith, Jaime G. Ramirez, and Vaclav Repa, A New Biology: A Modern Perspective on the Challenge of Closing the Gap between the Islands of Knowledge: in M. Cezon and Y. Wolfsthal (Eds.): ServiceWave 2011 Workshops, LNCS 6569, pp. 188--195. Springer, Heidelberg (2011).
M. Newton and L.S. Smith, Using spiking onset neurons and a recurrent neural network for musical sound classification.: poster presented at 161st Meeting of the Acoustical Society of America, Seattle, 23-27 May 2011. Abstract is at J. Acoustical Soc America, 129, 4 part 2, p2486. But there's a lot more (and more up-to-date) information on the poster itself.
R. Latif, E. Mastropaulo, A. Bunting, R. Cheung, T. Koickal, A. Hamilton, M. Newton, L. Smith Microelectromechanical systems for biomimetical applications: J. Vac. Sci. Technol. B 28(6), Nov/Dec 2010, DOI information: 10.1116/1.3504892
I.A. Gheyas and L.S. Smith, A neural network-based framework for the reconstruction of incomplete data sets: Neurocomputing, 73, 3039-3065. Published online 9 September 2010, DOI information: 10.1016/j.neucom.2010.06.021
L.S. Smith, Neuromorphic Systems: past, present and future: in Brain Inspired Cognitive Systems, A. Hussain, I. Aleksander, L.S. Smith, A.K. Barros, R. Chrisley, V. Cutsurdisis (eds), Springer Advances in Experimental Medicine and Biology 657, 2010, pp 167-182.
S. Shahid, J. Walker and L.S. Smith A new spike detection algorithm for extracellular neural recordings: IEEE Transactions on Biomedical Engineering, 57(4), 853-866, April 2010.
I.A. Gheyas, L.S. Smith, Feature subset selection in large dimensionality domains. Also available: Final draft of paper on Stirling Online Research Repository.: Pattern Recognition, 43, 1, 5-13, January 2010.
L.S. Smith and S. Shahid, Assessing the effectiveness of Cepstrum of Bispectrum based spike detection on simultaneously recorded intra- and extra- cellularly recorded data: presented at Society for Neuroscience Meeting, Chicago, 14-17 October 2009.
S. Shahid and L.S. Smith Cepstrum of Bispectrum Spike Detection applied to Extracellular Signals with Concurrent Intracellular Signals: Presented at CNS 2009 Berlin, July 2009
I.A. Gheyas, L.S. Smith A Neural Network Approach to Time Series Forecasting: presented at the 2009 International Conference of Computational Statistics and Data Engineering (ICCSDE), part of the World Congress in Engineering 2009 (London).
I.A. Gheyas, L.S. Smith A Novel Nonparametric Multiple Imputation Algorithm for Estimating Missing Data: presented at the 2009 International Conference of Computational Statistics and Data Engineering (ICCSDE), part of the World Congress in Engineering 2009 (London)
S. Shahid, L.S. Smith, Extracellular spike detection using Cepstrum of Bispectrum (powerpoint): presented at Society for Neuroscience Meeting, Washington DC, USA, November 2008.
M. Fletcher, B. Liang, L.S. Smith, A. Knowles, T. Jackson, K. Jessop, J. Austin, Neural network based pattern matching and spike detection tools and services - in the CARMEN neuroinformatics project: Neural Networks, 21, 8, 1076-1084, 2008.
S. Shahid and L.S. Smith (2008) A Novel Technique for Spike Detection in Extracellular Neurophysiological Recordings using Cepstrum of Bispectrum: presented at EUSIPCO 2008, 16th European Signal Processing Conference, August 25-29 2008.
J. Huo, A.F. Murray, L.S. Smith, Z. Yang (2008) Adaptation of Barn Owl Localization System with Spike Timing Dependent Plasticity: WCCI 2008/2008 International Joint Conference on Neural Networks IEEE Catalog Number: CFP08IJS-CDR ISBN: 978-1-4244-1821-3, ISSN: 1098-7576, June 1-6, 2008, Hong Kong. pages 155-160, 2008.
Giacomo Indiveri, R. Douglas, L.S. Smith (2008) Silicon neurons. Scholarpedia, 3(3):1887: Silicon neurons article in scholarpedia
Leslie S. Smith Artificial general intelligence: an organism and level based position statement: position paper accepted for the Artificial General Intelligence Conference (AGI-08), Memphis, March 2008.
Leslie S. Smith, Shahjahan Shahid (University of Stirling, UK), Anthony Vernier (Univ. de franche comte,France) Testing spike detection techniques using synthetic data: Poster presented at Society for Neuroscience Meeting, November 2007, San Diego
L.S. Smith, Neuronal computing or computational neuroscience: brains vs. computers: Seminar in Computational Thinking Series, Department of Informatics, Edinburgh University, October 17 2007.
L. S. Smith, S. Shahid, A. Vernier, N. Mtetwa Finding events in noisy signals: The IET Irish Signals and Systems Conference 13-14 September 2007, 31-36, 2007 (ISBN 978 0 86341 847 1).
L.S. Smith, J. Austin, S. Baker, R. Borisyuk, S. Eglen, J. Feng, K. Gurney, T. Jackson, M. Kaiser, P. Overton, S. Panzeri, R. Quian Quiroga, S.R. Schultz, E. Sernagor, V.A. Smith, T.V. Smulders, L. Stuart, M. Whittington, C. Ingram. The CARMEN e-Science pilot project: Neuroinformatics work packages.: Proceedings of the UK e-Science All Hands Meeting 2007 10-13 September 2007 (ed: S. J. Cox), 591-598, 2007 (ISBN 978-0-9553988-3-4).
Jenny Ure 1 , Rob Procter, Maryann Martone, David Porteous, Sharon Lloyd, Stephen Lawrie, Dominic Job, Richard Baldock, Alistair Philp, Dave Liewald, Frank Rakebrandt, Alan Blaikie, Clare McKay, Stuart Anderson, John Ainsworth, Jano van Hemert, Ignacio Blanquer, Richard Sinnott, Christian Barillot, Frank Bernard Gibaud, Alan Williams, Mark Hartswood, Paul Watson, Leslie Smith, Albert Burger, Jessie Kennedy, Horacio Gonzalez-Velez, Robert Stevens, Oscar Corcho, Robin Morton, Pamela Linksted, Mylene Deschenes, Mark McGilchrist, Paul Johnson, Alex Voss, Renate Gertz, Joanna Wardlaw, Data integration in eHealth: a domain/disease specific roadmap: Studies in Health Technol Inform, 2007;126:144-53.
L.S. Smith, S. Collins Determining ITDs using two microphones on a flat panel during onset intervals with a biologically inspired spike based technique: IEEE Transactions of Audio, Speech and Language Processing, 15, 8, 2278-2286, (2007). The stimuli used may be found here.
L.S. Smith, N. Mtetwa, A tool for synthesizing spike trains with realistic interference: Journal of Neuroscience Methods 159 (2007) 170-180
L.S. Smith Building a synthetic sensory-motor system: Originally presented at UK CRC Grand Challenges meeting, Edinburgh 2002, and updated in 2006. An over-the-top brief paper about extending life to new environments.
L.S. Smith, Neuroinformatics: what can E-Science offer Neuroscience (Or E-Science and Neuroscience: Experimental, Computational and Cognitive): Keynote presentation at Brain Inspired Cognitive Systems 2006 (BICS 2006), Molyvos, Lesvos, Greece, October 2006.
N. Mtetwa, L.S. Smith, Smoothing and thresholding in neuronal spike detection: Neurocomputing, 69, 10-12, pp 1366-1370, 2006
L.S. Smith Implementing Neural Models in Silicon: Handbook of Nature-Inspired and Innovative Computing: Integrating Classical Models with Emerging Technologies, ed A. Zomaya, Springer US, 2006, pp433-475
C.S. Thomas , C.A. Howie and L.S. Smith, A New Singly Connected Network Classifier based on Mutual Information, Intelligent Data Analysis. Abstract: Intelligent Data Analysis, Volume 9, Number 2, pp 189-205, 2005
Leslie Smith and Dagmar Fraser, Onsets, autocorrelation functions and spikes for direction based source separation (.pdf of powerpoint slides) (Abstract (JASA, 117, 4, p2485)): Presented at ASA conference, Vancouver, May 2005.
N. Mtetwa and L.S. Smith, Precision constrained stochastic resonance in a feedforward neural network.. Abstract and reference data.: IEEE Trans. Neural Networks, 16, 1, pp 250-262, 2005.
K. Parussel and L.S. Smith, Cost minimisation and Reward maximisation. A neuromodulating minimal disturbance system using anti-hebbian spike timing-dependent plasticity,: in Proceedings of the symposium on agents that want and like, motivational and emotional roots of cognition and action, pp 98-101, AISB 2005, 12-15 April 2005, ISBN 1 902956 41 7.
L.S. Smith, Towards Robot Audition: in Dynamic Perception, U.J. Ilg, H. Buelthoff, A Mallot (editors), IOS press/infix, 15-20, 2004.
L.S. Smith, D. S. Fraser, Robust sound onset detection using leaky integrate and fire neurons with depressing synapses,: IEEE Transactions on Neural Networks, 15, 5, (Sept 2004), pp 1125-1134.
Sound Signal Statistics,: Leslie S Smith, poster presented at Gordon Research Conference, Oxford, September 2004. I ought to try to turn it into a paper.
Sound feature detection using leaky integrate-and-fire neurons, Leslie S. Smith and Dagmar S. Fraser
Neuron/Electronic Interfacing: (warning: huge (13Mbyte) file) seminar presentation: Seminar presented at Bioengineering Center, Georgia Institute of Technology, Atlanta, Georgia, USA, 23 April 2003.
Biologically inspired robust onset detection (abstract), L.S. Smith,: Journal of the Acoustical Society of America, 113, 4 (Part 2), p2198, April 2003.
Presentation from above meeting
Stochastic resonance and finite resolutions in a network of leaky integrate and fire neurons, N. Mtetwa, L.S. Smith, A. Hussain,: Artificial Neural Networks - ICANN 2002, edited by J. R. Dorronsoro, Lecture Notes in Computer Science 2415, Springer 2002, pp 117-122.
Phase-locked onset detectors for monaural sound grouping and binaural direction finding (abstract), L.S. Smith: Journal of the Acoustical Society of America, 111, 5 (Part 2), p2467, May 2002.
Using IIDs to estimate sound source direction L.S. Smith: in From animals to animats 7, (eds: B. Hallam, D. Floreano, J. Hallam, G. Hayes, J-A Meyer), MIT Press, pp60-61, 2002.
Analogue VLSI Leaky Integrated-and-fire Neurons and their Use in a Sound Analysis System, Analog Integrated Circuits and Signal Processing, M.Glover, A.Hamilton L.S.Smith,: Special Issue: Microelectronics for Bio-inspired Systems (Selected Papers from MicroNeuro'99 Conference), Guest Editors: Alberto Prieto and Andreas Andreou, 30(2): 91-100 Feb 2002.
Using depressing synapses for phase locked auditory onset detection, L.S. Smith: in Artificial neural Networks - ICANN 2001, edited by G. Dorffner, H. Bischof, K. Hornik, Lecture Notes in Computer Science 2130, Springer, 2001.
Method and apparatus for processing sound L.S. Smith (inventor): World patent WO 00/001200 (WO 00001200), published 6 January 2000, (abandoned). International Patent Classification H04R 25/00, A1.
C. Breslin and L. S. Smith, Silicon Cellular Morphology: International Journal of Neural Systems, 9, 5, (Special Issue on Neuromorphic Systems), pp491-495, 1999
A Comparison of a Hardware and a Software Integrate and Fire Neural Network for Clustering Onsets in Cochlear Filtered Sound (CRC), L.S. Smith, M.A, Glover, A. Hamilton.: Accepted (poster) for Workshop on Neural Networks for Signal Processing, Aug 31-Sept 3, Cambridge, 1998: published in Neural Networks for Signal Processing VIII, Proceedings of the 1998 Workshop, edited by T. Constantinides, S. Y. Kung, M. Niranjan, E. Wilson, IEEE cat 98th8378.
Onset clustering (a mechanism for sound segmentation) uses integrate-and-fire neurons to perform across spectrum and across time clustering of increases in sound intensity in different parts of the spectrum. We show that a network of recently developed analogue VLSI integrate-and-fire neurons can perform this task in real-time, and compare its performance with a simulated network.
A One-dimensional Frequency Map Implemented using a Network of Integrate-and-fire Neurons (CRC), L.S. Smith: Accepted (poster) for ICANN 98, Skovde , September 2-4 1998. Published in ICANN 98: Proceedings of the 8th International Conference on Artificial Neural Networks, Skovde, Sweden, 2-4 September 1998, Springer, Perspectives in Neural Computing Series, Volume 2, p991-996, ISBN 3 540 76293 9.
A network of integrate-and-fire units (consisting of five excitatory units and one inhibitory unit) is shown to implement a one dimensional frequency map over one octave (80 - 160Hz). The network has a biologically plausible structure, conforming to Dale's law, and using plausible synaptic and axonic timings.
Reinforcement Landmark Learning P.Toombs, W.A. Phillips, L.S. Smith: Accepted for SAB 98, 17-21 August, Zurich, Switzerland
An analog VLSI integrate-and-fire neural network for sound segmentation (CRC), M.A. Glover, A. Hamilton, and L.S. Smith: Accepted for NC 98, Vienna 23-25 September 1998
This paper presents a cascadable aVLSI integrate-and-fire neural network chip (SPIKE I) capable of realistic biological time constants incorporated into a real time software based sound segmentation system with results. The sound segmentation system is based on an engineering abstraction of the functionality of the cochlea and auditory nerve. A comparison of the software simulation and software/hardware combination results indicates that clustering does occur. Furthermore the patterns of onsets and offsets generated are broadly similar. Analysis of the results indicates area's for improvement. These have been included in a second integrate-and-fire neural network chip (SPIKE II) presently being fabricated.
Adding lateral inhibition to a simple feedforward network enables it to perform exclusive--or. Smith L.S.: Neural Computation, 10, 2, 277-280, 1998
A simple laterally inhibited recurrent network which implements exclusive--or is demonstrated. The network consists of two mutually inhibitory units with logistic output function each receiving one external input, and each connected to a simple threshold output unit. The mutually inhibitory units settle into a point attractor. We investigate the range of steepness of the logistic, and the range of inhibitory weights for which the network can perform exclusive--or.
A Noise-robust Auditory Modelling Front End for Voiced Speech Smith L.S.: in Gerstner W., Germond A., Hasler M., Nicoud J-D (eds) Artificial Neural Networks - ICANN97, Lecture Notes in Computer Science 1327, pp 97-102, Springer-Verlag, Heidelberg, 1997, ISSN 0302-9743, ISBN 3-540-63631-5.
A method for detecting and displaying voiced elements of speech using amplitude modulated pulses due to unresolved harmonics of the excitation frequency (fundamental) is presented. It uses an auditory model consisting of a gammatone filterbank (modelling the basilar membrane), simple rectification (modelling the organ of Corti inner hair cells), envelope bandpass filters (modelling some spiral ganglion neuron effects) and amplitude modulation detectors (modelling certain cell populations in the cochlear nucleus). We demonstrate that it can display a pattern of activity across the spectrum and across time that describes the energy distribution in voiced speech, and that this pattern degrades slowly in the presence of non-speech noise.
Extracting Features from the Short-term Time Structure of Cochlear Filtered SoundSmith L.S.: in: John A. Bullinaria, David W. Glasspool & George Houghton (1998). Proceedings of the Fourth Neural Computation and Psychology Workshop: London, 9-11 April 1997 Connectionist Representations, pp113-125. London: Springer-Verlag.
Auditory modelling uses the architecture of the auditory system to guide early sound processing. The advantage of this approach is (i) time-resolution is better and (ii) many bandpassed channels are available and can be processed in parallel. Good time-resolution allows sophisticated across-time processing to be applied to each channel, resulting in the discovery of features in each channel. Logically each channel can be processed simultaneously. The features discovered can be correlated across channels. We present some early results for processing sound at three different levels of short-term time structure.
Using a Framework to Specify a Network of Temporal NeuronsSmith L.S.: Paper presented at 1st Slovak Symposium on Neural Networks and their Applications, November 11-13 1996, Herlany, Slovakia.
We discuss the use of frameworks (or formal models) for networks of temporal neurons, that is, neural networks using neurons in which precise signal timing matters. After discussing why one might require a framework at all, we review existing frameworks, and discuss the limitations of existing frameworks for their application to this more general form of neural network. With the aid of an example (a recurrent network of integrate-and-fire neurons) we show how one framework can be applied to this general form of neural network.
A neurally motivated technique for voicing detection in speech. (poster abstract)Smith L.S.: Poster presented at British Society of Audiology short papers meeting, Cambridge, England, September 1996. Published as Smith L.S., A neurally motivated technique for voicing detection in speech, (abstract), British Journal of Audiology, Vol 31, No 2, p112, 1997.
Envelope amplitude modulation occurs in cochlear filtered speech because of unresolved harmonics. The abstract is a brief introduction to the CCCN Technical Report 22, July 1996.
A Neurally Motivated Technique for Voicing Detection and F₀ Estimation for Speech.Smith L.S.: CCCN Tech Report 22, July 1996.
Speech consists of alternating voiced and unvoiced sections. Voiced speech consists of multiple harmonics of some fundamental ($F_{0}$); unvoiced speech consists of silence, or filtered noise. Here, speech is wideband bandpass filtered into many bands (modelling the cochlea). Each filter output is rectified (modelling the organ of Corti hair cell action), and bandpass filtered by convolution with the difference between two Gaussian averaging functions. This detects and emphasises the amplitude modulation resulting from unresolved harmonics (and models the combined effect of the auditory nerve and certain cochlear nucleus cell types). This output is compressed, summed across the bands, then used to discover glottal pulses. The presence of glottal pulses signals voicing, and the time between glottal pulses is used to find $F_{0}$. Results show good performance, particularly on male speakers. The system is reasonably resistant to background noise.
Using an Onset-based Representation for Sound SegmentationSmith L.S.: Using an onset-based representation for sound segmentation, p 274-281, Neural networks and their applications, Marseilles, March 20-22, 1996.
We present a technique for using pre-processing based on mammalian early auditory processing to produce a segmentation of sound based on onsets and offsets. The sound signal is bandpassed and each band processed to enhance onsets and offsets. The onset and offset signals are compressed, then clustered both in time and across frequency channels using a network of integrate-and-fire neurons. A spike-based representation of onsets and offsets is produced, and the timing of these spikes used to segment the sound. By considering spikes in varying number of bands, a multi-level segmentation tree can be built. This tree is a purely data-driven representation of the segmental structure of the sound.
Onset-based Sound SegmentationSmith L.S.: Onset based sound segmentation, pp 729--735, in Touretzky D.S., Mozer M.C., Hasselmo M.E. (eds) Advances in Neural Information Processing Systems 8 (Proceedings of the 1995 Conference), MIT Press, 1996.
A technique for segmenting sounds using processing based on mammalian early auditory processing is presented. The technique is based on features in sound which neuron spike recording suggests are detected in the cochlear nucleus. The sound signal is bandpassed and each signal processed to enhance onsets and offsets. The onset and offset signals are compressed, then clustered both in time and across frequency channels using a network of integrate-and-fire neurons. Onsets and offsets are signalled by spikes, and the timing of these spikes used to segment the sound.
A simple model of amplitude modulation detection. (poster abstract)Smith L.S.: Poster presented at British Society of Audiology short papers meeting, Oxford, England, September 1995. Published as Smith L.S., A simple model of amplitude modulation detection, (abstract) , British Journal of Audiology, 30, 2, 1996.
Deterioration of hearing at high frequencies leads to problems in speech interpretation in noise. One candidate for a carrier of useful information at higher frequencies is amplitude modulation (AM) found in wideband bandpassed voiced speech due to unresolved F0 harmonics. Taking an approach based in auditory modelling, we seek to identify (and eventually characterise) voiced sounds.
Synchronization of Integrate-and-fire Neurons with Delayed Inhibitory Lateral Connections.Smith L.S., Nischwitz A., Cairns D.E.: Synchronization of integrate-and-fire neurons with delayed inhibitory lateral connections, pp142-145, Proceedings of ICANN94, edited by M.Marinaro and P.G.Morasso, Springer-Verlag, 1994.
Integrate-and-fire (leaky integrator) neurons are both mathematically tractable and have a degree of biological plausibility. Systems of two neurons, interacting via symmetric pulsatile coupling with zero delay and zero absolute refractory period have been studied by Mirollo and Strogatz. For positive coupling, they found two fixed points, an unstable one with the units out of phase, and a stable one with the units in phase. For negative coupling, the stable and unstable fixed points are reversed, if a refractory period is assumed.
Sound Segmentation Using Onsets and Offsets, Smith L.S.: Journal of New Music Research, 23, 1, 11-23, March 1994.
Neural Networks, Free Association, and Errors, Smith L.S.: Poster summary, from AISB 1993.
A speculative discussion of the form a neural network might have if it was to display some of the characteristics Freud describes in his Psychology of Errors, and of free association.
Position Paper: Processing sound for interpretation.: CCCN Technical Report CCCN-17, September 1993
By taking an ecological view of sound perception strongly influenced by J.J. Gibson, we consider what the sound itself can usefully tell us about the source. We are therefore interested in extracting appropriate stimulus information. Since we also believe that low-level human auditory processing is independent of the nature of the sound, be it speech or not, we are interested in how the low-level human auditory system extracts such information. The eventual aim of this work is synthetic sound interpretation systems which can perform the simple tasks which we take for granted as humans. A brief description of a programme of research is included.
A Framework for Neural Net Specification, Smith L.S.: IEEE Transactions on Software Engineering, 18(7), 601-612, July 1992. Paper describing a notation for the specification of neural nets.
A Hardware Random Number Generator for Transputer Systems, Smith L.S., Kelly F.,: Occam User Group Newsletter, No 14, 43-45, January 1991. Short paper describing a hardware random number generator.

Back to Professor Smith's home page.

Last updated: Tuesday, 18-May-2021 16:50:20 BST

If you have any difficulties accessing this page, or you have any queries/suggestions arising from this page, please email:
Prof Leslie S Smith (lss(nospam_please)@cs.stir.ac.uk)