Kev tiv thaiv ntiag tug ntsuas thaum tsim cov ntaub ntawv hluavtaws
Thaum synthesizing ib dataset, nws yog ib qho tseem ceeb uas cov ntaub ntawv hluavtaws tuav tsis muaj cov ntaub ntawv rhiab heev uas tuaj yeem siv los txheeb xyuas cov tib neeg. Txoj kev no, peb tuaj yeem lav tias tsis muaj PII hauv cov ntaub ntawv hluavtaws. Hauv daim vis dis aus hauv qab no, Marijn qhia txog kev ntsuas kev ceev ntiag tug uas nyob hauv peb daim ntawv tshaj tawm zoo los qhia qhov no.
Cov vis dis aus no tau raug ntes los ntawm Syntho x SAS D[N]A Café txog AI Generated Synthetic Data. Nrhiav cov video tag nrho ntawm no.
Cov kev tiv thaiv kev ceev ntiag tug peb ua dab tsi thaum tsim cov ntaub ntawv hluavtaws?
Feem ntau, cov no yog metrics los tiv thaiv overfitting, saib nyob deb-kev ntsuas. Qhov no txhais tau tias lawv xyuas seb cov ntaub ntawv hluavtaws yog li cas rau cov ntaub ntawv qub. Yog tias ze dhau lawm, tej zaum yuav muaj kev pheej hmoo tsis pub lwm tus paub. Cov kev ntsuas no ua kom paub tseeb tias cov ntaub ntawv hluavtaws tsis tau ze rau cov ntaub ntawv qub. Tsis tas li ntawd, thaum ua qhov no, lub tshuab Syntho kuj tseem siv lub tshuab tuav kom muaj peev xwm ua tau qhov no hauv txoj kev ncaj ncees.
Cov ntaub ntawv yog hluavtaws, tab sis peb pab neeg yog tiag tiag!
Hu rau Syntho thiab ib tus ntawm peb cov kws tshaj lij yuav ntsib nrog koj ntawm qhov nrawm ntawm lub teeb los tshawb xyuas tus nqi ntawm cov ntaub ntawv hluavtaws!