1、Ha nds-On Reinfo rcement Lea rning with Py tho nMaster rein fo rcemen t an d d eep rein fo rcemen t learn in g u sin g Op en AI Gy man d Ten so rFlo wSu d h arsan Rav ich an d iranDo wn lo ad fro m fin ely b o o k 7 4 5 0 9 1 1 q q .co m2BIRMINGHAM - MUMBAIDo wn lo ad fro m fin ely b o o k 7 4 5 0 9
2、 1 1 q q .co m3Ha nds-On Reinfo rcement Lea rningwith Py tho nCo p y rig h t 2 0 1 8 Pack t Pu b lish in gAll rig h ts reserv ed . No p art o f th is b o o k may b e rep ro d u ced , sto red in a retriev al sy stem, o r tran smittedin an y fo rm o r b y an y mean s, with o u t th e p rio r written p
3、 ermissio n o f th e p u b lish er, ex cep t in th e caseo f b rief q u o tatio n s emb ed d ed in critical articles o r rev iews.Ev ery effo rt h as b een mad e in th e p rep aratio n o f th is b o o k to en su re th e accu racy o f th e in fo rmatio np resen ted . Ho wev er, th e in fo rmatio n co
4、 n tain ed in th is b o o k is so ld with o u t warran ty , eith er ex p ress o rimp lied . Neith er th e au th o r, n o r Pack t Pu b lish in g o r its d ealers an d d istrib u to rs, will b e h eld liab le fo ran y d amag es cau sed o r alleg ed to h av e b een cau sed d irectly o r in d irectly b
5、 y th is b o o k .Pack t Pu b lish in g h as en d eav o red to p ro v id e trad emark in fo rmatio n ab o u t all o f th e co mp an ies an dp ro d u cts men tio n ed in th is b o o k b y th e ap p ro p riate u se o f cap itals. Ho wev er, Pack t Pu b lish in g can n o tg u aran tee th e accu racy o
6、f th is in fo rmatio n .Co mmissio ning Edito r: Su n ith Sh ettyAcquisitio n Edito r: Namrata PatilCo ntent Dev elo pment Edito r: Amrita No ro n h aTechnica l Edito r: Jo v ita Alv aCo py Edito r: Safis Ed itin gPro ject Co o rdina to r: Sh weta H Birwatk arPro o frea der: Safis Ed itin gIndex er:
7、 Rek h a NairGra phics: Jish a Ch iray ilPro ductio n Co o rdina to r: Sh an tan u Zag ad eFirst p u b lish ed : Ju n e 2 0 1 8Pro d u ctio n referen ce: 1 2 6 0 6 1 8Pu b lish ed b y Pack t Pu b lish in g Ltd .Liv ery Place3 5 Liv ery StreetBirmin g h amB3 2 PB, UK.ISBN 9 7 8 -1 -7 8 8 8 3 -6 5 2 -
8、Do wn lo ad fro m fin ely b o o k 7 4 5 0 9 1 1 q q .co m4To my a d o ra b le p a ren ts, to my b ro th er, Ka rth ikeya n , a n d to my b estest frien d , Nikh il Ad itya . S u d h a rsa n Ra vich a n d ira nDo wn lo ad fro m fin ely b o o k 7 4 5 0 9 1 1 q q .co m5mapt.ioMap t is an o n lin e d ig
9、 ital lib rary th at g iv es y o u fu ll access to o v er 5 ,0 0 0 b o o k san d v id eo s, as well as in d u stry lead in g to o ls to h elp y o u p lan y o u r p erso n ald ev elo p men t an d ad v an ce y o u r career. Fo r mo re in fo rmatio n , p lease v isit o u rweb site.Do wn lo ad fro m fin
10、 ely b o o k 7 4 5 0 9 1 1 q q .co m6Why subscribe?Sp en d less time learn in g an d mo re time co d in g with p ractical eBo o k san d Vid eo s fro m o v er 4 ,0 0 0 in d u stry p ro fessio n alsImp ro v e y o u r learn in g with Sk ill Plan s b u ilt esp ecially fo r y o uGet a free eBo o k o r v
11、id eo ev ery mo n thMap t is fu lly search ab leCo p y an d p aste, p rin t, an d b o o k mark co n ten tDo wn lo ad fro m fin ely b o o k 7 4 5 0 9 1 1 q q .co m7Pa cktPub.co mDid y o u k n o w th at Pack t o ffers eBo o k v ersio n s o f ev ery b o o k p u b lish ed , withPDF an d ePu b files av a
12、ilab le? Yo u can u p g rad e to th e eBo o k v ersio n at www.PacktP an d as a p rin t b o o k cu sto mer, y o u are en titled to a d isco u n t o n th eeBo o k co p y . Get in to u ch with u s at fo r mo re d etails.At www.PacktP, y o u can also read a co llectio n o f free tech n ical articles,s
13、ig n u p fo r a ran g e o f free n ewsletters, an d receiv e ex clu siv e d isco u n ts an do ffers o n Pack t b o o k s an d eBo o k s.Do wn lo ad fro m fin ely b o o k 7 4 5 0 9 1 1 q q .co m8Co ntributo rsDo wn lo ad fro m fin ely b o o k 7 4 5 0 9 1 1 q q .co m9Abo ut the a utho rSudha rsa n Ra
14、v icha ndira n is a d ata scien tist, research er, artificialin tellig en ce en th u siast, an d Yo u Tu b er (search fo r S u d h a rsa n rein fo rcemen tlea rn in g ). He co mp leted h is b ach elo rs in in fo rmatio n tech n o lo g y at An n aUn iv ersity . His area o f research fo cu ses o n p r
15、actical imp lemen tatio n s o f d eeplearn in g an d rein fo rcemen t learn in g , wh ich in clu d es n atu ral lan g u ag ep ro cessin g an d co mp u ter v isio n . He u sed to b e a freelan ce web d ev elo p er an dd esig n er an d h as d esig n ed award -win n in g web sites. He is an o p en so u
16、 rceco n trib u to r an d lo v es an swerin g q u estio n s o n Stack Ov erflo w.I wo u ld like to th a n k my a ma zin g p a ren ts a n d my b ro th er, Ka rth ikeya n , fo r co n sta n tly in sp irin g a n dmo tiva tin g me th ro u g h o u t th is jo u rn ey. My b ig th a n ks a n d g ra titu d e to my b estest frien d , Nikh il Ad itya ,wh o is litera lly th e b estest, a n d to my ed ito r, Amrita , a n d to my S o eo r. With o u t a ll th eir su p p o rt, itwo u ld h a ve b een imp o ssib le to co mp lete th is b o o k.Do wn lo ad fro m fin ely b o o k 7 4 5 0 9 1 1 q q .co m1 0
