An episode can be any action or sequence of actions - how I lifted a cup, how I
had "tea" - a whole set of drinks and dishes, how I walked to the shops.
But you have to start with "your" actions, sequences of actions - and it is
only by laying down memories of these that you can make the eventual jump to
understanding the actions (and episodes) of others.
Work out, pace Koestler, what simple instruction can "trigger" a whole sequence
of moves/actions - and that instruction can be the core memory from which the
attached/consequent series of moves can be reconstructed when needed via
reflective memory.
Just a quick thought.
From: Piaget Modeler
Sent: Tuesday, July 17, 2012 12:22 AM
To: AGI
Subject: RE: [agi] Episodes
Let's forget about databases, scripts, and any other non-essential
representational formalisms.
So we assume the cognitive system is the mind of a robot, receiving a stream of
stimuli as frames
(visual, auditory, kinaesthetic) at moments in time (let's assume a moment is
3.15 miliseconds for
the sake of argument I have a practical reason for this as well).
And to avoid the "it's too complex" argument, let's simplify the problem by
assuming further that
only 1 bit of information comes in every moment per modality (i.e. 1 visual
bit, 1 auditory bit,
1 kinaesthetic bit, etc.).
This is terra artificialis incognita ...
What is an episode in your definition and how is it constructed? What should
happen after the bits arrive?
AGI | Archives | Modify Your Subscription
-------------------------------------------
AGI
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-c97d2393
Modify Your Subscription:
https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-2484a968
Powered by Listbox: http://www.listbox.com
