Ikibazo gihuza ibibazo · Hagati Urwego

Ikibazo Cyibisubizo Igisubizo

Ongera usubiremo ibibazo biboneka gusubiza, kubogama kwa dataset, guhuza multimodal no kugenzura imipaka kubishusho nibibazo byinyandiko.

Inzira zijyanye no kuyoboraIgisubizo Cyibibazo Igisubizo
Ikibazo 1 Bya 8

Ni ubuhe buryo bubiri sisitemu yo gusubiza ibibazo igaragara ifata?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Ni ubuhe buryo bubiri sisitemu yo gusubiza ibibazo igaragara ifata?
    • Ishusho nibibazo byururimi-karemano
    • Amashusho abiri yo kugereranya
    • Ikibazo na clip clip
    • Gusa ishusho idafite inyandiko
  2. Kuki dataset ya VQA v2.0 yashizweho hamwe nibisubizo 'byuzuye'?
    • Gukora amashusho murwego rwo hejuru
    • Kugabanya ururimi kubogama kuburyo moderi idashobora gukeka uhereye kumyandiko yonyine
    • Kuraho byose yego / ntakibazo
    • Kugabanya dataset yo guhugura byihuse
  3. Niki 'kubogama kururimi' muri VQA?
    • Guhindura ibibazo mururimi rutari rwo
    • Gukoresha indimi nyinshi icyarimwe
    • Gusubiza ukurikije amagambo yibibazo mugihe wirengagije ishusho
    • Guhitamo amashusho magufi kurenza ayandi
  4. Nigute moderi nyinshi za VQA zihuza amashusho nibibazo byamakuru?
    • Muguhuza pigiseli mbisi ninyuguti
    • Mugukora ishusho nshya uhereye kubibazo
    • Mu kwirengagiza ikibazo rwose
    • Muguhuza kodegisi zabo zerekana, akenshi hamwe no guhuza ibitekerezo
  5. Sisitemu ya VQA isanzwe itanga ibisubizo mukora iki?
    • Gutondekanya hejuru yamagambo ahamye yibisubizo bisanzwe
    • Gushushanya agasanduku gahuza gusa
    • Guhindura ishusho yumwimerere
    • Gusubiza ibyinjijwe ikibazo bidahindutse
  6. Niyihe ngero yikibazo cya VQA cyateye imbere?
    • Guhindura ishusho kuri thumbnail
    • Gusubiza ibibazo bijyanye ninyandiko nimibare imbere yishusho yinyandiko
    • Gutondekanya amashusho ukurikije izina rya filename
    • Gucomeka dosiye
  7. How does the original VQA benchmark score an answer when people phrase the same answer differently?
    • It requires an exact match to one fixed phrase only
    • It scores image resolution instead of answers
    • It compares the prediction with multiple human answers using a consensus-based score
    • It uses the model’s confidence as the ground truth
  8. Why can high VQA accuracy still fail to prove visual grounding?
    • Images cannot be represented by neural networks
    • Questions are always randomly generated
    • The benchmark contains no answer labels
    • A model may exploit dataset or language regularities instead of the image