Word Recognition and Learning based on Associative Memories and Hidden Markov Models

Authors: Zöhre Kara Kayikci, Günther Palm

Abstract:

A word recognition architecture based on a network of neural associative memories and hidden Markov models has been developed. The input stream, composed of subword-units like wordinternal triphones consisting of diphones and triphones, is provided to the network of neural associative memories by hidden Markov models. The word recognition network derives words from this input stream. The architecture has the ability to handle ambiguities on subword-unit level and is also able to add new words to the vocabulary during performance. The architecture is implemented to perform the word recognition task in a language processing system for understanding simple command sentences like “bot show apple".

Keywords: Hebbian learning, hidden Markov models, neuralassociative memories, word recognition.

Digital Object Identifier (DOI): doi.org/10.5281/zenodo.1055687

References:

[1] L. Rabiner and B. H. Juang, Fundemantal of Speech Recognition, Prentice-Hall, Inc., Upper Saddle River,, 1993.
[2] S. Young, et al., The HTK book for HTK version 3.2.1., Cambrige University, Engineering Department, 2002.
[3] G. Palm, "On associative memory," Biological Cybernetics, vol. 36, pp. 19-31, 1980.
[4] F. Schwenker, F. T. Sommer and G. Palm, "Iterative retrieval of sparsely coded associative memory patterns," Neural Networks, vol. 9(3), pp. 445-455, 1996.
[5] G. Palm, F. Kurfess, F. Schwenker and A. Strey, Neural Associative Memories. Technical Report, Universitat Ulm, Germany, 1993.
[6] A. Knoblauch and G. Palm, "Pattern separation and synchronization in spiking associative memories and visual areas," Neural Networks, vol. 14, pp. 763-780, 2001.
[7] H. Markert, A. Knoblauch and G. Palm, "Modeling of syntactical processing in the cortex," BioSystems, vol. 89, pp. 300-315, 2007.
[8] D. Willshaw, O. Buneman and H. Longuet-Higgins, "Non-holographic associative memory," Nature, vol. 222, pp. 960-962, 1969.
[9] G. Palm, "Memory capacities of local rules for synaptic modification. A comparative review," Concepts in Neuroscience, vol. 2, pp. 97-128, 1991.
[10] J. Buckingham and D. Willshaw, "Performance characteristics of associative nets," Network:Computation in Neural Systems, vol. 3, pp. 407-414, 1992.
[11] F. T. Sommer and G. Palm, "Improved bidirectional retrieval of sparse patterns stored by hebbian learning," Neural Networks, vol. 12, pp. 281- 297, 1999.
[12] D. O. Hebb, The Organization of Behaviour, John Wiley, Newyork, 1949.
[13] TIMIT Acoustic-Phonetic Continuous Speech Corpus, National Institute of Standartsand Technology Speech Dics 1-1.1, NTIS Order No. PB91- 505065, 1990.
[14] Z. Kara Kayikci, H. Markert and G. Palm, "Neural associative memories and hidden Markov models for speech recognition," presented at 2007 Int. Joint Conf. on Neural Networks, Orlando, Florida (USA).
[15] S. Young, ATK An Application Toolkit for HTK Version 1.6, Machine Intelligence Laboratory, Cambrige University, Engineering Department, 2007.
[16] Z. Kara Kayikci, D. Zaykovskiy, H. Markert, W. Minker and G. Palm Distributed Architecture for Speech Controlled Systems Based on Associative Memories Chapter in Mathematical Analysis of Evolution, Information and Complexity, Wiley-VCH, Weinheim (Germany), unpublished.