Abstract
We employ a special combination of different networks in order to process (transient) spatiotemporal patterns. In a first layer, feature analyzing cells translate instantaneous spatial patterns into activities of cells symbolizing the presence of certain feature values. A second layer maps the time sequence of symbols into a spatial activity pattern of the so-called TIM-cells. A third layer recognizes predefined activity patterns. We demonstrate the behaviour of the network using gaussian patterns in (1 + 1) space-time dimensions.
Similar content being viewed by others
References
Banzhaf W (1991) Processing spatio-temporal patterns by mapping time into intensity. In: Proc IJCNN Seattle 1991 11:871–877
Banzhaf W, Haken H (1990) Learning in a competitive network. Neural Networks 3:421–435
Cohen MA, Grossberg S (1987) Masking fields: a massively parallel architecture for learning, recognizing, and predicting multiple groupings of patterned data. Appl Opt 26:1866–1891
Haken H (1987) Synergetic computers for pattern recognition and associative memory. In: Haken H (ed) Computational systems, natural and artificial. Proceedings of the Elmau International Symposium on Synergetics 1987. Springer, Berlin Heidelberg New York, pp 1–21
Haken H (1988) Nonequilibrium phase transitions in pattern recognition and associative memory. Z Phys B-Condensed Matter 70:121–123
Hecht-Nielsen R (1987) Nearest matched filter classification of spatiotemporal patterns. Appl Opt 26:1892–1899
Nowlan SJ (1990) Maximum likelihood competitive learning. In: Touretzky DS (ed) Advances in neural information processing systems, vol 2. Morgan Kaufmann, San Mateo, Calif pp 574–582
Wang D, Arbib M (1990) Complex temporal pattern sequence learning based on short-term memory. Proc IEEE 78:1536–1543
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Banzhaf, W., Kyuma, K. The time-into-intensity-mapping network. Biol. Cybern. 66, 115–121 (1991). https://doi.org/10.1007/BF00243287
Received:
Accepted:
Issue Date:
DOI: https://doi.org/10.1007/BF00243287