quiz bu lëkkale ak tegtal · Diggante Niveau

Quiz jàng buy toppandoo

Saytu klonaasu jeffin, coppite ci séddale, fësal eksper, dajale done DAgger ak ay yamaleg ci biti réew yiñ seetlu.

Yooni njiit yu jëm ciJàng buy toppandoo
Laaj 1 de 8

Ci jàngu toppandoo, naka lay def ba ab ndawu liggéey di am jikkoom?

Xoolaatal laaj bu nekk8 laaj

Sooy def quiz bi, tontu yi ak leeral yi dañuy feeñ.

  1. Ci jàngu toppandoo, naka lay def ba ab ndawu liggéey di am jikkoom?
    • Jàng ab politik ci topp li eksper yi wane
    • Yokkateg neexal buñ defaree loxo jaaraleko ci seetlu bu bari
    • Komprime ab reso neuronal ngir dox ci ay aparey yu ndaw
    • Defar nataal ci laaj mbind
  2. Ban xeetu jafe-jafe mooy klonaasu jeffin?
    • Clustering buñu dul saytu
    • Jàngat buñ saytu ci kàrtograafi etaa-ci-jëf
    • jikkoy noon
    • Neexal jëmmal
  3. Ban jafe-jafe moo waral klonaasu doxalin bi baña mëna dem ci jëf yu yàgg yu toppalante?
    • Overfitting ci fonction neexal
    • Degrade yiy réer ci biir enkodeer bi
    • Jumtukaay buy yokk bu bawoo ci coppite ci séddale
    • Fatte bu metti klaas yu yàgg yi
  4. naka la algorithm DAgger di saafaraa ñakk kattan googu?
    • Soo dindi eksper bi ci loop bi yépp
    • Yokk facteur de réduction
    • Sooy yokk bruit ci jëf yépp
    • Soo laajee eksper bi ci réew yi jàngat bi di dem, ba noppi dajale done yooyu
  5. Lan moo waral jàng toppandoo di gëna taamu jàng dooleel ci liggéey yu melni dawal?
    • Demonstraasioŋ yi ñoo gëna yomba joxe, moo gën joxe neexal bu jaar yoon
    • Dafay garanti jeffin ju baax
    • Du soxla benn done
    • Dafay gëna gaaw ci inference kese
  6. What is a realistic limitation of behavioral cloning when the agent reaches states outside the demonstrations?
    • It is guaranteed to improve beyond the expert
    • Its behavior is not guaranteed there because those states may be out of distribution
    • It can only produce random actions
    • It automatically queries an expert at every step
  7. In one DAgger iteration, what does the learner do before requesting expert labels?
    • Deletes all earlier demonstrations
    • Trains only on randomly generated labels
    • Rolls out its current policy and records the states it actually visits
    • Replaces the expert policy permanently
  8. After the expert labels the states visited by the current policy, what happens to that data?
    • It is used only as an evaluation metric
    • It replaces the original observation space
    • It is discarded after one gradient step
    • It is aggregated with previous labeled data for retraining