MiuLab / DDQ

Deep Dyna-Q: Integrating Planning for Task-Completion Dialogue Policy Learning
MIT License
150 stars 44 forks source link

what is pre-training dqn model and world model ? initialize Q(s; a; θQ) and M(s; a; θM) via pre-training on human conversational data? #8

Open netrookiecn opened 4 years ago

netrookiecn commented 4 years ago

Hi I dont understand the pretraining of the world model because I can not find the pretraining process in your code, can you explain me what is that? and where is the pretraining dqn model and world model in your repo? thanks

Dr-Corgi commented 4 years ago

I have the same question. lol