Guide-yakabatana mibvunzo · Pakati Nhanho

Domain Adaptation Quiz

Yedza kwaunobva uye kuchinjika kwazvinonangwa, kudzoreredzwa kwedomasi gradients, kushomeka kwechinangwa-chinyorwa, nzira dzemahara-yemahara, uye kupatsanurwa kwezvinangwa.

Nzira dzekutungamira dzinoenderanaKugadzirisa Domaini
Mubvunzo 1 ye 8

Nderipi dambudziko rakagadzirirwa kugadzirisa domain?

Tarisa mubvunzo wese8 mibvunzo

Mhinduro uye tsananguro zvinoonekwa paunenge uchitora bvunzo.

  1. Nderipi dambudziko rakagadzirirwa kugadzirisa domain?
    • Kuderedza nhamba yemuenzaniso parameters
    • Kumhanyisa kuwedzera matrix paGPUs
    • Kuita kunodonha kana data reyedzo richisiyana nedata rekudzidzisa
    • Kugadzira zvinyorwa zvekudzidzira zvekugadzira otomatiki
  2. Mune domain-adversarial network, chii chinoitwa negradient reversal layer?
    • Inosundidzira chinhu chinodonhedza kuti chiite kuti sosi uye chinangwa chisazivikanwe
    • Iyo inobvisa iyo domain classifier panguva yekufungidzira
    • Inopeta kaviri mwero wekudzidza wekukurumidza kudzidziswa
    • Inoshandura regression outputs kuita classification labels
  3. Chii chinoita kuti 'usupervised' domain adaptation inonyanya kubatsira?
    • Izvo hazvidi kudzidziswa kwemuenzaniso zvachose
    • Inoshanda chete pane data data
    • Inobvisa kudiwa kwenzvimbo yenzvimbo
    • Iyo haidi data yakanyorwa mune inotangwa domain
  4. Ndeipi mamiriro emuenzaniso wekare we domain shift?
    • Modhi yakadzidziswa uye yakaedzwa pane iyo chaiyo dataset
    • A tumor classifier akadzidziswa pane imwe scanner achitadza pane imwe scanner
    • A modhi inomhanya nekukurumidza mushure me quantization
    • Network ine ma layer akawanda kupfuura imwe network
  5. Chii chinonzi 'source-free' domain adaptation nzira yakagadzirirwa kudzivirira?
    • Kuwana iyo yekutanga sosi yekudzidziswa data panguva yekuchinja
    • Kushandisa chero neural network architecture
    • Kugadzira chero fungidziro yekubuda
    • Kumhanya pamaGPU emazuva ano
  6. A team starts adaptation from a broadly pretrained model. What should it still measure?
    • Only the parameter count
    • Only source training accuracy
    • Performance on the actual target data and task
    • Whether target evaluation can be skipped
  7. Which DANN head supplies the objective to distinguish source examples from target examples?
    • The task label vocabulary
    • The domain classifier
    • The optimizer checkpoint file
    • A fixed output embedding dimension
  8. Does gradient reversal mean changing a positive class label into a negative class label?
    • Yes; every task label is negated
    • Yes; all target samples are relabeled as source
    • No; it freezes every feature weight
    • No; it reverses the domain gradient into the feature extractor