3 ms·
> Statistical learning can typically be phrased in terms of k nearest neighbours We have suspected that neural nets are a kind of kNN. Here's a paper: Every M
by visarga 3y ago
> Statistical learning can typically be phrased in terms of k nearest neighbours
We have suspected that neural nets are a kind of kNN. Here's a paper:
Every Model Learned by Gradient Descent Is Approximately a Kernel Machine
https://arxiv.org/abs/2012.00152 https://arxiv.org/abs/2012.00152