Direct gradient-based reinforcement learning.

Jonathan Baxter, Peter L. Bartlett

Browse the full ISCAS paper archive.