Reward-Weighted Regression Converges to a Global Optimum.

Miroslav Strupl, Francesco Faccio, Dylan R. Ashley, Rupesh Kumar Srivastava, Jrgen Schmidhuber

Browse the full AAAI paper archive.