TY  - RPRT
TI  - An Approximate Newton Method for Markov Decision Processes
AU  - Thomas Furmston
AU  - David Barber
PY  - 2015
UR  - https://arxiv.org/abs/1204.1227
ID  - 1204.1227
ER  -