On Sunday, June 27, 2021, at 2:06 PM, James Bowery wrote:
> I'd had the idea of a "truth speaker" for several years before that, but it 
> hadn't crystalized in my mind as a lossless compression competition until I 
> came up with the idea of what I called "The C-Prize" and Matt Mahoney 
> referred me to Marcus Hutter's work.

I remember reading some of this back then since I had been working on 
compression in the early 90's and had stopped but have followed it on and off 
since. When I pursued compression, I explicitly avoided probabilistic 
approaches being a contrarian as I knew everyone going to pursue that. Just 
like now with AI I avoid neural nets since, everyone is pursuing that, even 
though I was into NN's in the 90's.

Lossless compression and truth, hmmm, fake news will still have the fake 
crunched down. If you have enough news, you can denoise the fakeness as 
anomalistic... or how about lossless compress the fake news and have it 
losslessly decompress into estimated true news + some chunk of junk data. It’s 
still lossless but another representation. You can have a multitargeted 
lossless decompression there is nothing wrong with that.

Are you saying the K complexity of data is the ultimate truth? What's your idea 
of truth?
------------------------------------------
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/Tabec1b959b565c28-Mf52f27eb8aba2f4a16213b39
Delivery options: https://agi.topicbox.com/groups/agi/subscription

Reply via email to