Maswali ya Kujifunza ya Kuiga

Majibu na maelezo huonekana unapojibu maswali.

1. In imitation learning, how does an agent acquire its behavior?
   • Kujifunza sera kwa kunakili maonyesho ya wataalam
   • Kuongeza zawadi iliyoundwa kwa mikono kupitia uvumbuzi wa nasibu
   • Inabana mtandao wa neva ili kuendesha kwenye vifaa vidogo
   • Inazalisha picha kutoka kwa vidokezo vya maandishi
2. Uigaji wa fremu za tabia za kujifunza kama ni aina gani ya tatizo?
   • Mkusanyiko usiosimamiwa
   • Mafunzo yanayosimamiwa ya upangaji wa hali kwa hatua
   • Kizazi cha adui
   • Uundaji wa tuzo
3. Ni tatizo gani linalosababisha uundaji wa kitabia kushindwa katika mlolongo mrefu wa hatua?
   • Kutoshea kupita kiasi kwa utendaji wa zawadi
   • Taratibu zinazopotea kwenye kisimbaji
   • Kuchanganya hitilafu kutoka kwa zamu ya usambazaji
   • Usahaulifu mbaya wa madarasa ya zamani
4. Algorithm ya DAgger inashughulikiaje udhaifu huo?
   • Kwa kuondoa mtaalam kutoka kwa kitanzi kabisa
   • Kwa kuongeza kipengele cha punguzo
   • Kwa kuongeza kelele nasibu kwa vitendo vyote
   • Kwa kumuuliza mtaalamu wa majimbo mwanafunzi hutembelea na kujumlisha data hiyo
5. Kwa nini ujifunzaji wa kuiga mara nyingi hupendelewa kuliko kujifunza kwa kuimarisha kwa kazi kama vile kuendesha gari?
   • Maonyesho ni rahisi kutoa kuliko utendaji sahihi wa zawadi
   • Inahakikisha tabia bora
   • Haihitaji data yoyote
   • Inaendesha haraka kwa ufahamu tu
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