Quiz aktër-kritik

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

1. Ci anamu Actor-Critic, lan mooy wareefu 'critique' bi?
   • Xayma valeur u etaa yi (wala jëf yi) ak jàngat tànneefi aktër bi
   • Ngir tànn jëf ji nga wara jël
   • Ngir denc tampon replay
   • Ngir banneexu ci environmaa bi
2. Lan mooy 'njariñ' A(s,a) = Q(s,a) - V(s)?
   • Neexal episod bi yépp
   • Ñaata jëf moo gëna baax ci valeur moyenne bu etaa bi
   • Tolluwaayu jàng
   • Limu jëf yi jàppandi
3. Lan moo waral yokk benn ŋàññikat di wàññi faraas bi buñu ko méngale ak pexe yu sell yu melni REINFORCE?
   • Dina dindi neexal yi ci equation bi
   • Dafay tax politik bi nekk luy dogal
   • Critique bi dafay joxe ab baseline, suko defee yeesali jëfandikoo siñaal bu gëna néew bruit
   • Dafay yokk facteur de réduction
4. Ban kàttan la pexe Actor-Critic mëna am ci pexe yu lalu ci valeur yu melni Q-Learning ñu mëna jëfandikoo ci anam wu jaxasoo?
   • Jëf yu wuute
   • Jëfandikoo neexal
   • Wàññi ëlëg
   • Barabu jëf yuy wéy
5. Ci yii, yan ci ñoom mooy algorithm bu sukkandiko ci Actor-Critic te ñuy jëfandikoo bu bari ci RLHF ci misaali làkk yi?
   • PPO (Optimisation ci wàllu politik)
   • K-dafay tekki
   • Regresioŋ ligneer
   • Q-Jàng
6. Ban siñaal la aktër bi di faral di jëfandikoo ngir yeesal ay sàrtam?
   • Duggal pixel bu ñor bi
   • Njariñ wala njuumte TD bi ŋàññikat bi joxe
   • Dayo tampon biy baamtu
   • Nimero aleatoire bu fiks
7. An action has a positive advantage estimate. How does a policy-gradient actor typically use that signal?
   • Decrease its probability in every state
   • Set the critic’s value estimate to zero
   • Increase the probability of choosing that action in the state
   • Replace the policy with a Q-table
8. An action’s estimated advantage is negative. What direction should the actor update favor?
   • Increase it because the critic was surprised
   • Delete the state from the environment
   • Raise the reward discount factor automatically
   • Reduce the likelihood of that action relative to the policy baseline