While you used all sorts of uncommon confabulated words, I believe I already
"got" all of that. I believe there are specifics to the AGI tree/net that yous
know that I don't know, and that's what I want from you. Q-learning may as
well, I wish someone could explain it in few, concise words clearly. As far as
I'm aware, external reward like food, genital, body acceleration, etc, reward
senses when felt, just like frequency does in PPM (text prediction). And
internal reward is just a static force for predicting certain domains over
otherslike in Blender/ PPLM, but is also learnable if used ex.
word2vec/seq2vec. As for the secret details in Q_learning, first of all I think
it's for external body RL if I recall, not NLP? Now, external RL is very
limited, it's just rewarding actions close in time to the trigger, and updates
a hierarchy of motor programs, while external 'lab discovery tests' involve
collecting specif data, not reward, and so is more of a sensory thing than
motor thing. The AGI is searching for data/ patterns, the learnt motor
hierarchy is only to help it do so, it controls nothing...but what you want it
to, ignoring the few reflexes installed in us.
------------------------------------------
Artificial General Intelligence List: AGI
Permalink:
https://agi.topicbox.com/groups/agi/T0291e6571222d56c-M718fe314607d9e56b666d149
Delivery options: https://agi.topicbox.com/groups/agi/subscription
