Between Order and Chaos: The Quest for Meaningful Information Export

@article{citeulike:4062749, abstract = {Abstract\ \ The notion of meaningful information seems to be associated with the sweet spot between order and chaos. This form of meaningfulness of information, which is primarily what science is interested in, is not captured by both Shannon information and Kolmogorov complexity. In this paper I develop a theoretical framework that can be seen as a first approximation to a study of meaningful information. In this context I introduce the notion of facticity of a data set. I discuss the relation between thermodynamics and algorithmic complexity theory in the context of this problem. I prove that, under adequate measurement conditions, the free energy of a system in the world is associated with the randomness deficiency of a data set with observations about this system. These insights suggest an explanation of the efficiency of human intelligence in terms of helpful distributions. Finally I give a critical discussion of Schmidhuber's views specifically his notion of low complexity art, I\ defend the view that artists optimize facticity instead.}, author = {Adriaans, Pieter}, citeulike-article-id = {4062749}, citeulike-linkout-0 = {http://dx.doi.org/10.1007/s00224-009-9173-y}, citeulike-linkout-1 = {http://www.springerlink.com/content/635k27p66632646g}, day = {1}, doi = {10.1007/s00224-009-9173-y}, journal = {Theory of Computing Systems}, keywords = {information\_theory, learning\_theory, machine\_learning, mdl, thermodynamics}, month = {November}, number = {4}, pages = {650--674}, posted-at = {2009-08-31 22:42:39}, priority = {2}, title = {Between Order and Chaos: The Quest for Meaningful Information}, url = {http://dx.doi.org/10.1007/s00224-009-9173-y}, volume = {45}, year = {2009} }

TY - JOUR ID - citeulike:4062749 L3 - citeulike-article-id:4062749 N2 - Abstract  The notion of meaningful information seems to be associated with the sweet spot between order and chaos. This form of meaningfulness of information, which is primarily what science is interested in, is not captured by both Shannon information and Kolmogorov complexity. In this paper I develop a theoretical framework that can be seen as a first approximation to a study of meaningful information. In this context I introduce the notion of facticity of a data set. I discuss the relation between thermodynamics and algorithmic complexity theory in the context of this problem. I prove that, under adequate measurement conditions, the free energy of a system in the world is associated with the randomness deficiency of a data set with observations about this system. These insights suggest an explanation of the efficiency of human intelligence in terms of helpful distributions. Finally I give a critical discussion of Schmidhuber’s views specifically his notion of low complexity art, I defend the view that artists optimize facticity instead. IS - 4 JF - Theory of Computing Systems EP - 674 TI - Between Order and Chaos: The Quest for Meaningful Information VL - 45 SP - 650 KW - information_theory KW - learning_theory KW - machine_learning KW - mdl KW - thermodynamics AU - Adriaans, Pieter PY - 2009/11/01/ UR - http://dx.doi.org/10.1007/s00224-009-9173-y ER -