I have updated my universal intelligence test with benchmarks on about 100 compression programs.
http://cs.fit.edu/~mmahoney/compression/uiq/ The results seem to show good correlation with real data. The best compressors on this synthetic data are also the best on most benchmarks with real data. Although my goal was to sample a Solomonoff distribution to measure universal intelligence (as defined by Hutter and Legg), this is not possible because we would have to sample from halting machines determined by coin flips, and we don't know which ones halt. In practice, we can approximate a solution by placing time bounds on the machines. I think this is justified because real data sources have to be computable. I have also made some technical changes to the machine language in the data generating program. My goal was to use a simple but expressive language to output self delimiting strings in a format that most simple compressors could already model. -- Matt Mahoney, [email protected] ------------------------------------------- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244&id_secret=123753653-47f84b Powered by Listbox: http://www.listbox.com
