DocumentCode :
2187852
Title :
On the relation between descriptional complexity and algorithmic probability
Author :
Gács, Péter
fYear :
1981
fDate :
28-30 Oct. 1981
Firstpage :
296
Lastpage :
303
Abstract :
Several results in Algorithmic Information Theory establish upper bounds on the difference between descriptional complexity and the logarithm of "apriori probability". It was conjectured that these two quantities coincide to within an additive constant. Here, we disprove this conjecture and show that the known overall upper bound on the difference is exact. The proof uses a memory-allocation game between two players called User and Server. User sends incremental requests of memory space for certain structured items, Server allocates this space in a write-once memory. For each item, some of the allocated space is required to be in one piece, in order to live a short address. We also present some related results.
Keywords :
Approximation algorithms; Binary sequences; Computer science; Encoding; Entropy; Inference algorithms; Information theory; Probability distribution; Upper bound; Writing;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Foundations of Computer Science, 1981. SFCS '81. 22nd Annual Symposium on
Conference_Location :
Nashville, TN, USA
ISSN :
0272-5428
Type :
conf
DOI :
10.1109/SFCS.1981.31
Filename :
4568347
Link To Document :
بازگشت