Hacker News new | ask | show | jobs
Every Model Learned by Gradient Descent Is Approximately a Kernel Machine (medium.com)
13 points by mathematically 1696 days ago