AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
The Problem With Forward Chaining
This relates to an algorithm that's used in mathematical optimization, which is the idea of gradient ascent for going up or if you're going down. It's basically the idea of wherever you're standing right now, you say, which one little step can I take that goes up the fastest? Like I'm trying to get this up a mountain. Okay, I'll take one little step in that whatever direction goes up the very fastest. Then I'll look around again and figure out now in my new position what little step goes up the fastest. And you can prove that for many types of mountains, so to speak, this algorithm will eventually get you to the top of the mountain.