TY  - RPRT
TI  - Near-Optimal Glimpse Sequences for Improved Hard Attention Neural Network Training
AU  - William Harvey
AU  - Michael Teng
AU  - Frank Wood
PY  - 2020
UR  - https://arxiv.org/abs/1906.05462
ID  - 1906.05462
ER  -