Inspiration: Why is loud music much louder after pausing and resuming it?

The human ear includes the ability to adjust the volume of incoming sound by contracting some muscles. If the sound is too loud, the ear adjusts to make it quieter. If the sound is too quiet, the ear adjusts to make it louder. A similar function is performed by the human eyeball's pupils.

The general concept is that a complex processing model outputs some feedback to guide its own input pre-processing stages in the next round.

Has any notable machine learning model ever used this concept to improve its accuracy? Is there a name for this concept?

  • $\begingroup$ FYI We are fundamentally different to machines in this aspect: we can't feel state, but change (first derivative). In general, all our senses can't perceive state, only change. Fixing the eyes over an image makes the image disappear. Same for touch, smell, taste, hearing. And this applies not only to physical senses, but also to metaphysical: thermoception, proprioception, etc. Anyway, of course there are learning models which profit of change (e.g. a video sliding window to detect certain behaviors). $\endgroup$
    – RodolfoAP
    Commented Jun 23, 2023 at 7:53


You must log in to answer this question.

Browse other questions tagged .