Abstract
"The least biased inference, taking available information into account, is the one with maximum entropy". So we are taught by Jaynes. The many followers from a broad spectrum of the natural and social sciences point to the wisdom of this principle, the maximum entropy principle, MaxEnt. But "entropy" need not be tied only to classical entropy and thus to probabilistic thinking. In fact, the arguments found in Jaynes' writings and elsewhere can, as we shall attempt to demonstrate, profitably be revisited, elaborated and transformed to apply in a much more general abstract setting. The approach is based on game theoretical thinking. Philosophical considerations dealing with notions of cognition - basically truth and belief - lie behind. Quantitative elements are introduced via a concept of description effort. An interpretation of Tsallis Entropy is indicated.
Original language | English |
---|---|
Article number | 040002 |
Journal | A I P Conference Proceedings Series |
Volume | 1853 |
Issue number | 1 |
Number of pages | 2 |
ISSN | 0094-243X |
DOIs | |
Publication status | Published - 9 Jun 2017 |