InoteveraGaidhi rinotevera
Kuronga Kwechiyero Chekudzidza
Tekinoroji
Nhungamiro yehunyanzvi
Q-Kudzidza
Q-Kudzidza ndeyekusimbisa kudzidza algorithm inodzidzisa mumiriri kuti ndeapi zviito zvinobhadhara zvakanyanya nekudzidza zvishoma nezvishoma kukosha kwekufamba kwega kwega kuburikidza nekuedza uye kukanganisa.
Teerera kugwaro iriAI-yakarondedzerwa · ~2 min
Pa peji ino2 min verenga
Pfupiso
It matters because it can find optimal behavior without ever being told the rules of its environment.
Kudzika Kwakadzika
Q-Kudzidza inodzidza basa rinodaidzwa kuti Q(s, a): mubairo unotarisirwa wenguva yakareba wekutora chiito 'a' muhurumende 's' uye wozoita zvakanaka pashure. Mumiririri anotanga kuziva chinhu, anoedza zviito, uye anoona mibairo. Mushure menhanho yega yega inokwenya fungidziro yayo yeQ-kukosha kumubairo uchangobva kugamuchirwa pamwe neramangwana rakaderedzwa remangwana kukosha kwainotarisira kubva kune inotevera nyika. Zvakanyanya kukosha, ndeye 'isina-policy' uye 'yemahara-yemhando': inogona kudzidza yakanakisa mutemo uchiongorora zvisina tsarukano, uye haimboda modhi yemashanduko enyika. Tichipihwa kuongororwa kwakaringana kwechero nyika-chiito pair, iwo maQ-values anochinjika kune akanyanya kukosha, uye chakanakisa chiito mune chero nyika ingori iyo ine yepamusoro Q.
Kuitwa muNyika yeChokwadi
Atari mutambo-kutamba vamiririri (DeepMind's DQN) kudzidza kutamba Breakout uye Pong zvakananga kubva pachiratidziri pixels.
Kunatsiridza traffic-mwenje nguva pamharadzano kuderedza yakazara mota yekumirira nguva
Robhoti navigation kuburikidza negridi kana maze uko iro robhoti rinodzidza ipfupi nzira yekuwedzera mubairo
Dynamic mitengo uye sarudzo dzekutsvaga apo mumiririri anodzidza kuti ndezvipi zviito zvinowedzera purofiti yenguva refu
Ramangwana reQ-Kudzidza
Classic tabular Q-Kudzidza zvinonetsa kana nyika dzakawandisa kuti dzichengete mutafura. Iyo inotungamira kutungamira iri kuisanganisa neneural network, sepakadzika Q-Networks (DQN), iyo inofananidzira Q-tsika kubva kune yakabikwa mapikisi. Tsvagiridzo inoenderera mberi nekudzikamisa izvi nekudzokororwa kweruzivo, network yakanangana, uye akasiyana seDouble DQN uye kugovera Q-Kudzidza kunoderedza fungidziro yakarerekera uye inomiririra kudzoreredzwa kwakazara kwete maavhareji.
Kuzvishandisa
Nei zvichikosha
- Mutengo uye bhajeti. Zvisarudzo zvezvivakwa zvinotyaira kuita uye mutengo wekushandisa kwemakore.
- Sarudzo dzakajeka. Dzidzo yehunyanzvi inobatsira zvikwata kusarudza murwi wakakodzera, kwete iwo mutsva chete.
- Kudzora kwemhando yepamusoro. Sarudzo dzeinjiniya dziri nani dzinoderedza zviitiko zvekuvimbika mukugadzira.
Chenjera
- Kugadzirisa imwe bhenji kunogona kuvanza yakafara system kushaya simba.
- Infrastructure uye mari yekugadzirisa inowanzotarisirwa pasi.
- Chengetedzo uye kucherechedzwa mapundu anogona kukura sezvo masisitimu anowedzera kuoma.
Matanho anotevera
- Tsanangura latency, mhando, uye mutengo zvinangwa usati waitwa.
- Benchmark pasi pechokwadi mutoro uye data mamiriro.
- Chishandiso chekutarisa zvikanganiso, kudonha, uye mushandisi maitiro.
- Gadzirira nzira dzekudzosera kumashure uye dzezviitiko usati wawedzera.
Ramba Uchiongorora
Tsamba yemahara
Tora iyo yemazuva ese AI pfupiso
Nhatu dzakasimbiswa dzeAI nyaya mangwanani ega ega pakati pevhiki, dzakanyorwa neChirungu chakajeka. Yemahara zvachose, hapana ads.
Imwe email zuva rega rega revhiki. Bvisa kunyoresa kamwe chete. Hatimbotengese kana kugovera kero yako.
Zviedze iwe pachako
Take the Q-Learning quiz
Mhinduro yepakarepo pamhinduro yega yega, uye chitupa chinogovaniswa chine ID inobatika kana wapedza kosi.
Tsigira dzidzo yemahara yeAI. AI Understanding ndeye 501(c)(3) isingabatsiri - hapana ads, hapana paywall, zvachose. Ita chipo
Mibvunzo inowanzo bvunzwa
What is Q-Learning?
Q-Kudzidza ndeyekusimbisa kudzidza algorithm inodzidzisa mumiriri kuti ndeapi zviito zvinobhadhara zvakanyanya nekudzidza zvishoma nezvishoma kukosha kwekufamba kwega kwega kuburikidza nekuedza uye kukanganisa. Izvo zvine basa nekuti inogona kuwana hunhu hwakakwana pasina kumboudzwa mitemo yenzvimbo yayo.
Ko Q-ukoshi Q(s, a) inomiririrei?
Q(s, a) inofungidzira mubairo wakaderedzwa weramangwana kubva pakuita chiito a in state s uye kuzvibata zvakanaka mushure mazvo, kwete mubairo wepakarepo.
Sei Q-Kudzidza ichitsanangurwa se 'off-policy'?
Iwo max pamusoro pezvinotevera zviito zvinoreva kuti Q-Kudzidza inodzidza kukosha kweanokara yakakwana mutemo kunyangwe mumiririri achiongorora neimwe nzira yekuzvibata.
Mumutemo wekuvandudza, chii chinodzikisira chinhu gamma chinodzora?
Gamma (pakati pe0 ne1) inoderedza mibairo yeramangwana; ma values ari padhuze na1 anoita kuti mumiririri aone zviri kure, values padhuze na 0 anoita kuti aone zvishoma.
Chii chinonzi temporal-difference (TD) kukanganisa muQ-Kudzidza?
Iko kukanganisa kweTD ndiyo mukaha pakati pefungidziro nyowani (mubairo pamwe nepamusoro wakadzikiswa kukosha kweramangwana) uye fungidziro yekare yeQ; iyo yekuvandudza inoderedza gap iri.
Sei yakajeka tabular Q-Kudzidza ichinetseka nematambudziko makuru semitambo yemavhidhiyo kubva kumapikisheni?
Tafura yekutarisa inoda yekupinda pa-state-action pair, izvo zvisingagoneke kana nyika dzichiwanda mumabhiriyoni, ichikurudzira neural-network approximators seDQN.
Ramba uchidzidza
Mazano anoenderana
Mamwe madhairekitori akasarudzirwa nyaya iyi