Previous |  Up |  Next


exponential family; information divergence
This article studies exponential families $\mathcal{E}$ on finite sets such that the information divergence $D(P\|\mathcal{E})$ of an arbitrary probability distribution from $\mathcal{E}$ is bounded by some constant $D>0$. A particular class of low-dimensional exponential families that have low values of $D$ can be obtained from partitions of the state space. The main results concern optimality properties of these partition exponential families. The case where $D=\log(2)$ is studied in detail. This case is special, because if $D<\log(2)$, then $\mathcal{E}$ contains all probability measures with full support.
[1] Ay, N.: An information-geometric approach to a theory of pragmatic structuring. Ann. Probab. 30 (2002), 416-436. DOI 10.1214/aop/1020107773 | MR 1894113 | Zbl 1010.62007
[2] Ay, N.: Locality of global stochastic interaction in directed acyclic networks. Neural Computat. 14 (2002), 2959-2980. DOI 10.1162/089976602760805368 | Zbl 1079.68582
[3] Brown, L.: Fundamentals of Statistical Exponential Families: With Applications in Statistical Decision Theory. Institute of Mathematical Statistics, Hayworth 1986. MR 0882001 | Zbl 0685.62002
[4] Cover, T., Thomas, J.: Elements of Information Theory. First edition. Wiley, 1991. MR 1122806
[5] Csiszár, I., Shields, P.: Information Theory and Statistics: A Tutorial. First edition. Foundations and Trends in Communications and Information Theory. Now Publishers, 2004.
[6] Csiszár, I., Matúš, F.: Generalized maximum likelihood extimates for exponential families. Probab. Theory Rel. Fields 141 (2008), 213-246. DOI 10.1007/s00440-007-0084-z | MR 2372970
[7] Pietra, S. Della, Pietra, V. Della, Lafferty, J.: Inducing features of random fields. IEEE Trans. Pattern Analysis and Machine Intelligence 19 (1997), 380-393. DOI 10.1109/34.588021
[8] Drton, M., Sturmfels, B., Sullivant, S.: Lectures on algebraic statistics. In: Oberwolfach Seminars 39, Birkhäuser, Basel 2009. MR 2723140 | Zbl 1166.13001
[9] Geiger, D., Meek, C., Sturmfels, B.: On the toric algebra of graphical models. Ann. Statist. 34 (2006), 5, 1463-1492. DOI 10.1214/009053606000000263 | MR 2278364 | Zbl 1104.60007
[10] Jaynes, E. T.: Information theory and statistical mechanics. Phys. Rev. 106 (1957), 4, 620-630. DOI 10.1103/PhysRev.106.620 | MR 0087305 | Zbl 0084.43701
[11] Juríček, J.: Maximization of information divergence from multinomial distributions. Acta Univ. Carolin. 52 (2011), 1, 27-35. MR 2808291
[12] Lauritzen, S. L.: Graphical Models. First edition. Oxford Statistical Science Series, Oxford University Press, 1996. MR 1419991
[13] Linsker, R.: Self-organization in a perceptual network. IEEE Computer 21 (1988), 105-117. DOI 10.1109/2.36
[14] Matúš, F., Ay, N.: On maximization of the information divergence from an exponential family. In: Proc. WUPES'03, University of Economics, Prague 2003, pp. 199-204.
[15] Matúš, F., Rauh, J.: Maximization of the information divergence from an exponential family and criticality. In: 2011 IEEE International Symposium on Information Theory Proceedings (ISIT2011), 2011.
[16] Montúfar, G., Rauh, J., Ay, N.: Expressive power and approximation errors of Restricted Boltzmann Machines. In: NIPS, 2011.
[17] Oxley, J.: Matroid Theory. First edition. Oxford University Press, New York 1992. MR 1207587
[18] Rauh, J.: Finding the Maximizers of the Information Divergence from an Exponential Family. Ph.D. Dissertation, Universität Leipzig, 2011. MR 2817016
[19] Rauh, J.: Finding the maximizers of the information divergence from an exponential family. IEEE Trans. Inform. Theory 57 (2011), 6, 3236-3247. DOI 10.1109/TIT.2011.2136230 | MR 2817016
[20] Rauh, J., Kahle, T., Ay, N.: Support sets of exponential families and oriented matroids. Internat. J. Approx. Reasoning 52 (2011), 5, 613-626. DOI 10.1016/j.ijar.2011.01.013 | MR 2787021
[21] Zhu, S. C., Wu, Y. N., Mumford, D.: Minimax entropy principle and its application to texture modeling. Neural Computation 9 (1997), 1627-1660. DOI 10.1162/neco.1997.9.8.1627
Partner of
EuDML logo