Ikibazo gihuza ibibazo · Biragoye Urwego

Ikibazo Cyimbitse cyo muri Bayesian

Shakisha imiyoboro ya Bayesian, uburemere butazwi, ibigereranyo byinyuma, Monte Carlo yataye ishuri, hamwe na kalibrasi yicyizere kumakuru yatanzwe.

Inzira zijyanye no kuyoboraKwiga Byimbitse
Ikibazo 1 Bya 8

Nigute urusobe rw'imitsi yo muri Bayesiya ifata uburemere bwarwo butandukanye nubusanzwe?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Nigute urusobe rw'imitsi yo muri Bayesiya ifata uburemere bwarwo butandukanye nubusanzwe?
    • Ifata buri buremere nkigabanywa rishoboka aho kuba agaciro kamwe gahamye
    • Ikoresha ibipimo byuzuye aho kuba decimals
    • Ikuraho uburemere rwose
    • Igabana uburemere bumwe murwego rwose
  2. What can a Bayesian approach provide that a single fixed-weight network does not directly represent?
    • A guarantee that every confidence score is calibrated
    • A distribution of predictions that reflects uncertainty about model weights
    • A proof that the model will detect every unfamiliar input
    • A way to remove all noise from the data
  3. How can epistemic uncertainty change as relevant, representative training data are added?
    • It is guaranteed to reach zero with any added data
    • It is identical to label noise
    • It may decrease under the chosen model and data assumptions
    • It always increases linearly
  4. Ni ukubera iki ibigereranyo nkibisobanuro bitandukanye cyangwa MCMC bikenewe?
    • Kuberako imiyoboro yimitsi idafite imikorere yigihombo
    • Kuberako GPU idashobora gukora inyongera
    • Kuberako ibipimo bigomba kuba byuzuye
    • Kuberako inyuma yukuri miriyoni yuburemere ntishobora kubarwa
  5. Niki guta ishuri rya Monte Carlo gukora kugirango ugereranye gushidikanya?
    • Komeza guta ishuri gukora mugihe cyibizamini kandi ikoresha umuyoboro inshuro nyinshi, ukoresheje ikwirakwizwa ryibisubizo
    • Hindura abata ishuri kandi imyitozo ndende
    • Kuraho neuron zose burundu
    • Gusimbuza igihombo n urusaku rudasanzwe
  6. Impinduka zinyuranye zisubiramo igereranya ryinyuma nkikibazo ki?
    • Ikibazo
    • Ikibazo cyo gutezimbere kigabanya gutandukana kwa KL (binyuze muri ELBO)
    • Ikibazo cyiza cyo gutoranya ikibazo kidafite intego
    • Ikibazo cyo gushushanya ibyuma
  7. How does Bayes’ rule relate a prior, likelihood, and posterior?
    • The posterior is the prior plus the likelihood, without normalization
    • It is the likelihood divided by the prior without normalization
    • The posterior is proportional to the prior times the likelihood
    • It is the fixed point estimate of the most likely weight vector
  8. How does MCMC approximate a Bayesian neural-network posterior?
    • By training a single point estimate with dropout disabled
    • By ranking model outputs with a fixed rule
    • By minimizing an ELBO with a simpler variational family
    • By drawing samples through a Markov chain whose stationary distribution targets the posterior