A Proposal for Reducing the Number of Trial-and-Error Searches for Deep Q-Networks Combined with Exploitation-Oriented Learning.

Naoki Kodama, Kazuteru Miyazaki, Taku Harada

Browse the full ICMLA paper archive.