The short version
A program-synthesis paper on wake-sleep library learning, singled out as a favorite.
From the HackerLinks archive
A program-synthesis paper on wake-sleep library learning, singled out as a favorite.
The short version
A program-synthesis paper on wake-sleep library learning, singled out as a favorite.
Why it caught our attention
An explicit favorite-paper endorsement offers a concrete research rabbit hole beyond the submitted story.
Where it surfaced on Hacker News
“Reminds me of one of my favourite ML papers, DreamCoder:”
macleginn · recommendation
Direct HN comment
Sleep framework: A new way for LLMs to retain knowledge in continual learning