non-stationarity of objectives
Problem where the rules or goals of optimization keep changing, preventing systems from converging on a fixed solution—a barrier to AI achieving genuine scientific discovery
As defined in The AI Revolution Is Underhyped | Eric Schmidt | TED
Where it comes up · 1
Most important claims · 2 of 2Search all claims for “non-stationarity of objectives” →
- Current AI systems cannot perform true scientific discovery because they can only recombine existing knowledge, but genuine breakthroughs (like Einstein's) require recognizing patterns across completely unrelated domains and transferring tools between them—a capability requiring solving the problem of 'non-stationarity of objectives' where the rules of discovery keep changing.The AI Revolution Is Underhyped | Eric Schmidt | TED
- Solving the non-stationarity problem—enabling AI systems to recognize when the rules of a problem are changing—will unlock the ability to perform true scientific discovery, but it requires systems to work on their own data-generation loops.The AI Revolution Is Underhyped | Eric Schmidt | TED