Guide-yakabatana mibvunzo · Pakati Nhanho

Multi-Task Kudzidza Quiz

Inotarisa kugovaniswa kwebasa rakaoma uye rakapfava, zvinangwa zvakayerwa, kutamisa kwakashata uye nzira dzekugadzirisa zvinopokana basa gradients.

Nzira dzekutungamira dzinoenderanaMulti Task Kudzidza
Mubvunzo 1 ye 8

How is a multi-task learning model set up during training?

Tarisa mubvunzo wese8 mibvunzo

Mhinduro uye tsananguro zvinoonekwa paunenge uchitora bvunzo.

  1. How is a multi-task learning model set up during training?
    • Kudzidzira modhi imwe pamabasa akati wandei ane hukama ane zvinomiririra zvakagovaniswa
    • Kudzidzira muenzaniso wakasiyana webasa rega rega
    • Kushandisa data isina kunyorwa chete yebasa rimwechete
    • Kubvisa ese akavanzika maseketi kubva kunetiweki
  2. Mune yakajairwa yakaoma-parameter-kugovera MTL network, chii chinogovaniswa uye chii chakaparadzana?
    • Zvese zvakapatsanurwa pabasa
    • Musana wakagovaniswa une basa-rakanangana nemisoro yekubuda
    • Chete basa rekurasikirwa rinogoverwa
    • Iyo chete data yekupinza, kwete maseru
  3. Sei kudzidza-mazhinji-basa kuchigona kuvandudza generalization?
    • Inobvisa kudiwa kwechero data
    • Inovimbisa zero kudzidziswa kukanganisa
    • Zviratidzo kubva kumabasa ane hukama anogadzirisa kumiririrwa kwakagovaniswa uye kuderedza kuwandisa
    • Inobvisa kudiwa kwemisoro yebasa
  4. Chii chinonzi 'negative transfer' mukudzidza-mabasa akawanda?
    • Kana kuwedzera mabasa kunomhanyisa kudzidziswa
    • Kana iyo modhi iine ma parameter mashoma
    • Kana mabasa ese achigovana mavara akafanana
    • Kana kudzidza kwerimwe basa kuchikuvadza kuita pane rimwe nekuda kwezviratidzo zvinopokana
  5. Iko kurasikirwa kwese kunowanzo kuumbwa sei mukudzidza-mabasa akawanda?
    • Sehuwandu hwakayerwa hwemunhu wega basa kurasikirwa
    • Nekutora basa rakakura chete kurasikirwa
    • Nokufuratira basa rimwe chete
    • Nekuwanza kurasikirwa pamwe chete
  6. Sei mhando dzemitauro mikuru ichionekwa seyakasiyana-siyana?
    • Vanogona kuita basa rimwe chete rakagadziriswa
    • Network imwe chete inobata kushandura, kupfupisa, kukodha, uye Q&A
    • Vanoshandisa modhi yakasiyana kune imwe neimwe kukurumidza
    • Ivo havambogovani ma parameter mumabasa ese
  7. In soft parameter sharing, how are task models usually related?
    • They use one identical model with no task-specific outputs
    • They share no information and cannot exchange features
    • They keep separate parameters but add a constraint or penalty encouraging similarity
    • They share only the training labels
  8. What does PCGrad do when two task gradients conflict?
    • It discards every task except the largest one
    • It changes the task labels
    • It replaces the shared backbone with separate datasets
    • It projects away conflicting gradient components before combining updates