Self-improving AI agents tend to memorize their test tasks, so their gains shrink or disappear on new ones. RRSI, a new method from Google researchers, reins in this effect and lifts scores on unseen benchmarks by up to 4.7 points while using about 30 percent fewer tokens than an unregularized version.
The article Google researchers find a way to keep self-improving AI agents from memorizing their tests appeared first on The Decoder.
source & further reading
the-decoder.com — original article
NASA and IBM's open source lunar model turns 17 years of orbiter data into a foundation for lunar science
Another OpenAI safety departure adds to a pattern of researchers leaving with public warnings
AI agents build 3D scenes from photos but have no idea if they got it right