How to learn the shape of a filter from only what it allowed through?

zakelfassi · x · 2026-08-26

The author poses an interesting thought: we can only train on data that passed through the 'judgment filter' and reached the public eye, not on what was excluded. Yet, those exclusions are where judgment actually happened. How do we learn the shape of a filter from only what it allowed through?

Original post →

More from Research

Research channel →