AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
The Problem Is, We're Learning Abstractions, Right?
The goal here is to figure out how to get a machine learning to learn structures that, by virtue of their simplification, are more generalizable. A lot ofthe abstractions, dit di mineral network learns, are like these low level, border line, spurious kinds of abstractions. That's why they break so easy. Like wind is an abstraction a good abstraction? I don't know, but i think it all kind of, in a way, misses the point.