On Thursday, November 18, 2021, at 12:15 PM, Matt Mahoney wrote:
> No, it's more lossy. Not that lossiness has a numeric value, but if it did, a 
> reasonable measure would be the percent reduction in size. The part we 
> discard typically has higher Kolmogorov complexity due to being mostly random 
> noise that can't be compressed losslessly.

Doesn’t it more depend on the data to be compressed, it’s complexity 
distribution, and how the two or more compressors work with each other over the 
data if we are talking about loss In regards to the whole, say if it’s an image.

An intent might be to introduce losslessness to lossyness to lose less overall 
instead of just better compression ratios. Though some regions of lossyness 
might become more lossy. And there are reasons in some cases to lose less even 
in the non-visualizable regions.

Across all scenarios of compressors, data, and interoperabilities - I'm not 
sure if it would be less lossy or more lossy or equal.
------------------------------------------
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/T5ff6237e11d945fb-Mbbbf1ea2ac63f33c7d92be84
Delivery options: https://agi.topicbox.com/groups/agi/subscription

Reply via email to