Ikibazo gihuza ibibazo · Hagati Urwego

Gukomeza Kwiga no Kwibagirwa Ikibazo

Ongera usubiremo kwibagirwa ibyago, gutuza-plastike, EWC, gusubiramo, imiryango ikora, hamwe nimbibi zo kwiga. Reba urugero n'ibitekerezo.

Ikibazo 1 Bya 8

'Kwibagirwa ibyago' ni iki?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. 'Kwibagirwa ibyago' ni iki?
    • Iyo wiga umurimo mushya utesha agaciro imikorere kumurimo wize mbere
    • Iyo icyitegererezo kibuze kwibuka kandi kigwa
    • Iyo amakuru yo guhugura asibwe kubwimpanuka
    • Iyo icyitegererezo cyanze gutanga ubuhanuzi
  2. Gucuruza kumutima wokwiga bikunze kwitwa:
    • Ibicuruzwa bitandukanye
    • Ikibazo gihamye
    • Umuvuduko wukuri
    • Ubushakashatsi-bukoreshwa mubucuruzi
  3. Niki Elastic Weight Consolidation (EWC) ikora?
    • Gusiba neuron idafite akamaro
    • Gukuba kabiri igipimo cyo kwiga kubikorwa bishya
    • Ihana impinduka kuburemere bwingenzi kubikorwa byize mbere
    • Ubike buri ngero zamahugurwa yashize
  4. Nigute uburyo bushingiye kubisubiramo burwanya kwibagirwa?
    • Muguhagarika umuyoboro wose burundu
    • Ukoresheje igipimo kinini cyo kwiga
    • Mugukuraho imikorere yigihombo
    • Muguhuza ububiko bwabitswe cyangwa bwatanze ingero zishaje mumahugurwa mashya
  5. How does Gradient Episodic Memory (GEM) constrain a gradient update while learning a new task?
    • It projects the gradient so its first-order approximation does not increase loss on stored prior-task examples
    • It replaces all prior-task data with new examples
    • It freezes every parameter for all future tasks
    • It increases weights that were unimportant to earlier tasks
  6. Ni ukubera iki naive-gutunganya neza umurimo mushya bitera kwibagirwa?
    • GPU irashyuha kandi ikangiza kwibuka
    • Gradient descente yimura uburemere busangiwe kumurimo mushya nta mbogamizi yo kubungabunga izishaje
    • Imirimo mishya burigihe ifite amakuru menshi
    • Icyitegererezo gisiba nkana ibiro bishaje
  7. How does iCaRL classify a test example after learning multiple classes?
    • By retrieving the single closest word in a text corpus
    • By selecting the class with the largest Fisher information
    • By comparing its feature vector with per-class means computed from stored exemplars
    • By replaying every training image at inference
  8. Why is evaluating continual learning more than measuring final accuracy on the newest task?
    • New-task accuracy is the only relevant measure
    • Old tasks are removed from the evaluation
    • The model cannot be tested on multiple tasks
    • A method must also measure retention on prior tasks and its ability to learn new ones