http://toohna.ourproject.org/RPG/index.php?action=history&feed=atom&title=AlMeta%3AYTDNM_reinforcement_learning 2026-05-04T23:23:51Z Revision history for this page on the wiki MediaWiki 1.30.2 http://toohna.ourproject.org/RPG/index.php?title=AlMeta:YTDNM_reinforcement_learning&diff=1781&oldid=prev 2022-06-28T08:55:51Z

<p>Created page with &quot;Your the Dog Now Ma(n|chine) ---- #if (you got a reward) ## if reward is &gt; 0 ###Treat =true ###add top of @History to @Trick ##else (if reward &lt; 0) ###Treat = False ###clear...&quot;</p> <p><b>New page</b></p><div>Your the Dog Now Ma(n|chine)<br /> <br /> ----<br /> #if (you got a reward)<br /> ## if reward is &gt; 0<br /> ###Treat =true<br /> ###add top of @History to @Trick<br /> ##else (if reward &lt; 0)<br /> ###Treat = False<br /> ###clear @Trick<br /> ##for X in @History<br /> ###update X in NN<br /> ###if (X was Conditioned) <br /> ####Reward = Reward minus( 1 minus likelihood of choosing X)<br /> ###if(X was not Conditioned)<br /> ####Reward = Reward times 0.5<br /> ##clear @History<br /> ##if (you are not sure what to do)<br /> ###if (Treat)<br /> ####if (second of @Trick not Conditioned)<br /> #####add @Trick to @Repertoire<br /> ###add new choice to @History as not Conditioned<br /> ##else (if you are sure what to do)<br /> ###add new choice to @History as Conditioned<br /> #else (if you didn't get a reward)<br /> ##if (your in not vary sure of what do)<br /> ###if (you last action was Conditioned)<br /> ####reward = -1<br /> ####while(top of @History was Conditioned)<br /> #####pop top of @History as X<br /> #####Update X in NN<br /> #####reward = reward times likelihood of X<br /> ####Clear @History<br /> ###Clear @Trick<br /> ###add new choice to @History as not Conditioned <br /> ##else (if you are sure of what to do)<br /> ###if you last action was Conditioned)<br /> ####add top of @History to @Trick <br /> ###add new choice to @History as Conditioned.</div>

Cmdrtako