Kediska Barashada Xoojinta gadaashiis

Jawaabaha iyo sharraxaadaha waxay soo baxayaan markaad qaadanayso kediska.

1. Maxay tahay barashada xoojinta gaddoonka ah ujeedadeedu tahay soo kabashada?
   • Shaqo abaal-marin ah oo sharraxaysa hab-dhaqanka khabiirka ee la arkay
   • Hagaajinta degdega ah ee soo degista gradient
   • Nooca mudaaharaadka
   • Astaanta xogta sawirka
2. Waa maxay sababta dhibaatada IRL loogu sifeeyay mid khaldan ama madmadow?
   • Sababtoo ah mudaaharaadyadu had iyo jeer waa kuwa ugu fiican
   • Sababtoo ah hawlo badan oo abaal-marineed oo kala duwan ayaa sharxi kara dabeecad isku mid ah
   • Sababtoo ah abaalgudyadu mar walba waa inay noqdaan kuwo taban
   • Sababtoo ah uma baahna xog khabiir
3. Waa maxay faa'iidada muhiimka ah ee soo kabashada abaalgudka ay ka leedahay cloning dhaqanka?
   • Uma baahna mudaaharaad
   • Had iyo jeer waxay ku socotaa waqti joogto ah
   • Waxa ay qabsata yoolka hoose, iyada oo guud ahaan dawladaha aan la arki karin
   • Waxay ka fogaataa adeegsiga siyaasad kasta
4. Sidee buu ugu badnaan-entropy IRL u xalliyaa madmadowga abaalmarinta?
   • Adigoo ku qasbaya dhammaan abaal-marinnada inay sinaadaan
   • Adigoo iska indhatiraya hab-raacyada hoose ee gebi ahaanba
   • Adiga oo meesha ka saaraya shaqada abaalmarinta
   • Iyada oo la moodayo itimaalka dhabta ah sida jibbaarada guud ahaan abaalgudka
5. Ka dib marka IRL soo kabsato shaqada abaal-marinta, maxaa caadi ahaan la sameeyaa marka xiga?
   • Barashada xoojinta caadiga ah waxay ka soo saartaa siyaasad abaalmarintaas
   • Abaalka waa la tuuray
   • Bandhigyada waa la tirtiray si loo badbaadiyo meel bannaan
   • Wakiilku wuxuu joojiyaa waxbarashada
6. Xaaladgee ayay IRL faa'iido gaar ah u leedahay?
   • Marka abaalgudku si fiican loo yaqaan
   • Marka aad rabto in aad fahanto oo aad guud u samayso yoolalka ka dambeeya hab-dhaqanka khabiirka
   • Marka aysan jirin xogta habdhaqanka
   • Marka aad u baahan tahay oo kaliya inaad xasuusato ficillada go'an
7. Why can a constant zero reward explain many different expert policies?
   • It assigns the highest reward only to expert actions
   • It proves the expert is optimal
   • Every policy receives the same return, so none is preferred by the reward
   • It makes the transition model deterministic
8. In maximum-entropy IRL, how are higher-reward trajectories treated relative to lower-reward ones?
   • They are assigned probability zero
   • They are treated as equally likely regardless of reward
   • They are removed before policy learning
   • They receive higher probability, with probability depending exponentially on total reward