Ikibazo gihuza ibibazo · Hagati Urwego

Gukuramo Isano Kuva Ikibazo

Gerageza uburyo gukuramo isano bihindura inyandiko mubintu byubatswe, nuburyo imiyoboro rusange hamwe nuburyo bwo kugenzura bishobora kunanirwa.

Inzira zijyanye no kuyoboraGukuramo Isano Kuva Inyandiko
Ikibazo 1 Bya 8

Nibihe bisohoka mubisanzwe sisitemu yo gukuramo isano?

Reba buri kibazo8 ibibazo

Ibisubizo nibisobanuro bigaragara mugihe ufata ikibazo.

  1. Nibihe bisohoka mubisanzwe sisitemu yo gukuramo isano?
    • Inzira eshatu zubatswe zihuza ibice bibiri nisano
    • Ikibonezamvugo parse igiti cyinteruro
    • Amanota yimyumvire kuva mubi kugeza nziza
    • Guhindura interuro mu rundi rurimi
  2. In a common pipeline for relation extraction, what may identify entity mentions before pair classification?
    • Guhindura inyandiko-ku-mvugo
    • Kumenyekanisha ibigo
    • Igice cy'ishusho
    • Audio source separation
  3. Why can distant supervision produce noisy relation-extraction labels?
    • Knowledge bases contain no structured facts
    • It requires every sentence to be hand-labeled
    • A sentence may mention two linked entities without stating their knowledge-base relation
    • It removes entity mentions before training
  4. Nigute moderi nyinshi za transformateur zishingiye kuri RE zerekana ibice byombi byo gutondekanya?
    • Muguhindura interuro mbere
    • Mugusiba kimwe mubice
    • Muguhindura inyandiko mumajwi
    • Mugushyiramo ibimenyetso byihariye byerekana ibimenyetso byombi
  5. Ni iki gitandukanya gukuramo 'gufungura' gukuramo RE bisanzwe?
    • Ikuramo interuro ifitanye isano itaziguye ninyandiko idafite urutonde rwateganijwe rwubwoko
    • Ikora ku mibare gusa
    • Irasaba amakuru yanditseho kurusha ubundi buryo ubwo aribwo bwose
    • Irashobora kubona isano imwe gusa kuri buri nyandiko
  6. Why include a no-relation label when classifying candidate entity pairs?
    • To translate the sentence into another language
    • To represent pairs for which the text does not assert a target relation
    • To replace named entity recognition
    • To guarantee that every pair has a relation
  7. What does document-level relation extraction allow a system to use?
    • Only the spelling of a single entity
    • Audio features from a spoken recording
    • Evidence that may be spread across multiple sentences
    • A fixed relation without reading text
  8. How does open relation extraction differ from a closed relation schema?
    • It cannot use text as input
    • It always requires a knowledge base to label every sentence
    • It only identifies entity boundaries
    • It can express relation phrases without choosing only from a predefined label inventory