×

zbMATH — the first resource for mathematics

Finite state automata resulting from temporal information maximization and a temporal learning rule. (English) Zbl 1090.68064
Summary: We extend Linkser’s Infomax principle for feedforward neural networks to a measure for stochastic interdependence that captures spatial and temporal signal properties in recurrent systems. This measure, stochastic interaction, quantifies the Kullback-Leibler divergence of a Markov chain from a product of split chains for the single unit processes. For unconstrained Markov chains, the maximization of stochastic interaction, also called Temporal Infomax, has been previously shown to result in almost deterministic dynamics. This letter considers Temporal Infomax on constrained Markov chains, where some of the units are clamped to prescribed stochastic processes providing input to the system. Temporal Infomax in that case leads to finite state automata, either completely deterministic or weakly nondeterministic.
Transitions between internal states of these systems are almost perfectly predictable given the complete current state and the input, but the activity of each single unit alone is virtually random. The results are demonstrated by means of computer simulations and confirmed analytically. It is furthermore shown numerically that Temporal Infomax leads to a high information flow from the input to internal units and that a simple temporal learning rule can approximately achieve the optimization of temporal interaction. We relate these results to experimental data concerning the correlation dynamics and functional connectivities observed in multiple electrode recordings.

MSC:
68Q45 Formal languages and automata
68T05 Learning and adaptive systems in artificial intelligence
PDF BibTeX XML Cite
Full Text: DOI
References:
[1] DOI: 10.1073/pnas.92.19.8616 · doi:10.1073/pnas.92.19.8616
[2] Abeles M., J. Neurophysiol. 70 pp 1629– (1993)
[3] Abeles M., Concepts in Neuroscience 4 pp 131– (1993)
[4] Aertsen A. M. H. J., J. Neurophysiol 61 pp 900– (1989)
[5] DOI: 10.1109/18.930911 · Zbl 0997.94009 · doi:10.1109/18.930911
[6] DOI: 10.1162/089976602760805368 · Zbl 1079.68582 · doi:10.1162/089976602760805368
[7] DOI: 10.1016/S0893-6080(03)00190-4 · Zbl 02060409 · doi:10.1016/S0893-6080(03)00190-4
[8] DOI: 10.1088/0954-898X/12/3/301 · doi:10.1088/0954-898X/12/3/301
[9] DOI: 10.1162/neco.1995.7.6.1129 · Zbl 05479838 · doi:10.1162/neco.1995.7.6.1129
[10] DOI: 10.1146/annurev.neuro.24.1.139 · doi:10.1146/annurev.neuro.24.1.139
[11] DOI: 10.1162/089976603321891774 · Zbl 1046.92011 · doi:10.1162/089976603321891774
[12] DOI: 10.1080/00401706.1994.10485840 · doi:10.1080/00401706.1994.10485840
[13] DOI: 10.1109/72.761705 · doi:10.1109/72.761705
[14] DOI: 10.1038/35090500 · doi:10.1038/35090500
[15] DOI: 10.1038/416433a · doi:10.1038/416433a
[16] DOI: 10.1088/0954-898X/8/3/005 · Zbl 0903.92007 · doi:10.1088/0954-898X/8/3/005
[17] DOI: 10.1109/10.16444 · doi:10.1109/10.16444
[18] Gütig R. G., Journal of Neuroscience 23 pp 3697– (2003)
[19] DOI: 10.1162/neco.1994.6.1.127 · doi:10.1162/neco.1994.6.1.127
[20] DOI: 10.1073/pnas.83.19.7508 · doi:10.1073/pnas.83.19.7508
[21] DOI: 10.1073/pnas.83.21.8390 · doi:10.1073/pnas.83.21.8390
[22] DOI: 10.1073/pnas.83.22.8779 · doi:10.1073/pnas.83.22.8779
[23] DOI: 10.1162/089976600300014872 · doi:10.1162/089976600300014872
[24] DOI: 10.1007/BF00199057 · Zbl 0826.92008 · doi:10.1007/BF00199057
[25] DOI: 10.1007/BF00199545 · Zbl 0840.92007 · doi:10.1007/BF00199545
[26] DOI: 10.1162/08997660260293238 · Zbl 1018.68068 · doi:10.1162/08997660260293238
[27] DOI: 10.1088/0954-898X/7/3/002 · Zbl 0898.92039 · doi:10.1088/0954-898X/7/3/002
[28] DOI: 10.1146/annurev.neuro.24.1.1193 · doi:10.1146/annurev.neuro.24.1.1193
[29] DOI: 10.1146/annurev.ne.18.030195.003011 · doi:10.1146/annurev.ne.18.030195.003011
[30] DOI: 10.1016/S0893-6080(00)00053-8 · doi:10.1016/S0893-6080(00)00053-8
[31] DOI: 10.1073/pnas.91.11.5033 · doi:10.1073/pnas.91.11.5033
[32] DOI: 10.1073/pnas.96.6.3257 · doi:10.1073/pnas.96.6.3257
[33] DOI: 10.1126/science.287.5456.1273 · doi:10.1126/science.287.5456.1273
[34] DOI: 10.1007/s12064-003-0033-z · doi:10.1007/s12064-003-0033-z
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.