Scaling Memory-Augmented Neural Networks with Sparse Reads and Writes (opens in new tab)

(xxx.lanl.gov)

3 pointsvisionscaper9y ago1 comments

1 comments

1 comments · 1 top-level

Sparse Access Memory is like a differentiable hash table. I'm wondering how many new differentiable data structures could be invented in order to augment deep neural nets?

j / k navigate · click thread line to collapse