Home

Awesome

Code for the article <a href='https://arxiv.org/abs/1705.09322'>Convergent Tree-Backup and Retrace with Function Approximation.</a>

<img src="https://github.com/ahmed-touati/convergent-off-policy/blob/master/plots/counterexample.png" title="2 state Counterexample where TB and Retrace diverge">

<img src="https://github.com/ahmed-touati/convergent-off-policy/blob/master/plots/counterexample_gradient.png" title="2 state Counterexample where GTB and GRetrace converge">