Proto-Predictive Representation of States with Simple Recurrent Temporal-Difference Networks thumbnail
Pause
Mute
Subtitles not available
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Proto-Predictive Representation of States with Simple Recurrent Temporal-Difference Networks

Published on Sep 17, 20092813 Views

We propose a new neural network architecture, called Simple Recurrent Temporal-Difference Networks (SR-TDNs), that learns to predict future observations in partially observable environments. SR-TDNs i