Martin, Would you recommend using policies? If so why? ~PM Date: Mon, 16 Mar 2015 01:59:44 +0000 Subject: Re: [agi] Plans vs. Policies From: [email protected] To: [email protected]
A policy is a one-step plan depending on observations/state. The (observation dependent) two-step plan is obtained from the policy by incorporating the next observation and performing the associated action. A mullti-step plan is then dependent on the multiple future observations. A plan that does not depend on future observations is a special case of this, and maybe is what AI planning does (but I don't know much about it). From: [email protected] To: [email protected] Subject: [agi] Plans vs. Policies Date: Sun, 15 Mar 2015 16:22:46 -0700 Reinforcement Learning uses "policies" to select actions while most work in AI Planning emphasizes the construction and representation of a "plan" which consists of a sequence of actions (or a hierarchyof composite and primitive actions). Kindly compare, contrast, evaluate trade-offs, and recommend the plans or policies approach Your rationale is appreciated. ~PM ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
