Ubuyobozi bwa tekiniki
Pseudo-Kwandika no Kwitoza
Pseudo-labeling ni tekinike igenzurwa na kimwe cya kabiri aho icyitegererezo cyatojwe kumurongo muto wanditseho ibyarango byacyo kubirango bitemewe, hanyuma bigatoza kuri ubwo buhanuzi.
Umva iki gitaboAI-byavuzwe · ~2 min
Kuriyi page2 min soma
Incamake
It is a simple, powerful way to exploit abundant unlabeled data.
Kwibira cyane
Kwimenyereza ni kimwe mu bitekerezo bya kera bigenzurwa. Urabanza guhugura icyitegererezo cya mwarimu kumibare ntarengwa yanditseho amakuru. Umwarimu noneho ahanura ibirango kuri pisine nini yingero zidashyizweho ikimenyetso; ibyiringiro byinshi guhanura bihinduka pseudo-labels. Umunyeshuri wicyitegererezo yatojwe kubumwe bwibirango nyabyo na pseudo-labels, akenshi biruta mwarimu. Icyizere ntarengwa ni ngombwa: gusa ibyahanuwe hejuru yo guhagarika ibishoboka birabikwa, bityo moderi ntabwo yangizwa nibitekerezo byayo bitazwi. Impinduka zigezweho zihuza pseudo-label hamwe nibisanzwe bihoraho. FixMatch, kurugero, itanga pseudo-label kuva ishusho yongerewe intege kandi igatoza icyitegererezo cyo kuyihuza na verisiyo yongerewe cyane, ariko mugihe gusa guhanura gukomeye kwizeye. Urusaku rwumunyeshuri rwagabanije igitekerezo kuri ImageNet muguhindura umunyeshuri no kongeramo urusaku (guta ishuri, kongera) mugihe cyamahugurwa.
Gushyira mu bikorwa Isi
Guhugura uburyo bwo kumenya imvugo wandukura amasaha ibihumbi byamajwi atanditseho urugero rwimbuto, hanyuma ukongera ukiga kumpapuro zizewe.
__AIU_PROTECTED_11 __ 'Umunyeshuri wuzuye urusaku atezimbere ImageNet yukuri yerekana ibimenyetso byerekana amashusho atanditse hamwe numwarimu no guhugura umunyeshuri munini, usakuza.
Kwandika pisine nini ya scan yubuvuzi itamenyekanye hamwe nicyitegererezo cyahuguwe kubibazo magana make byanditseho impuguke kugirango bagure amahugurwa.
Bootstrapping yinyandiko itondekanya kuri niche domaine ukoresheje pseudo-yerekana amamiriyoni yinyandiko zitanditse hejuru yicyizere.
Kazoza ka Pseudo-Kwandika no Kwitoza
Pseudo-labels ikomeza kuba intandaro yo kwiga-gukora neza kandi bigenda byiyongera kumiyoboro minini-yimyitozo ngororamubiri, aho moderi zikomeye zitanga ibirango byogukora cyangwa ndetse namakuru ya syntetique yo gutoza moderi ntoya cyangwa nshya, uburyo bwo gusibanganya. Tegereza kwishyira hamwe hamwe no kwiga gukora (guhitamo ingero abantu bagomba kuranga), igereranyo cyiza kidashidikanywaho cyo gushungura pseudo-labels, no gukomeza gukoreshwa mukumenyekanisha imvugo, amashusho yubuvuzi, hamwe na domaine iyo ari yo yose aho amakuru atanditseho umubare munini cyane wanditseho amakuru.
Kubishyira mubikorwa
Impamvu ari ngombwa
- Igiciro na bije. Ibyemezo byubwubatsi bitwara imikorere nigiciro cyimikorere kumyaka.
- Ibyemezo bisobanutse. Ubuhanga bwa tekinike bufasha amakipe guhitamo umurongo ukwiye, ntabwo ari shyashya gusa.
- Kugenzura ubuziranenge. Guhitamo neza bya injeniyeri bigabanya ibintu byizewe mubikorwa.
Witondere
- Gutezimbere igipimo kimwe gishobora guhisha intege nke za sisitemu.
- Ibikorwa Remezo no kubungabunga akenshi usanga bidahabwa agaciro.
- Icyuho cyumutekano no kwitegereza birashobora kwiyongera uko sisitemu igenda igorana.
Intambwe ikurikira
- Sobanura ubukererwe, ubuziranenge, nigiciro cyibiciro mbere yo kubishyira mubikorwa.
- Ibipimo byerekana umutwaro ufatika hamwe namakuru yimiterere.
- Gukurikirana ibikoresho kubikosa, drift, ningaruka zabakoresha.
- Tegura inzira yo gusubiza ibyabaye mbere yo gupima.
Komeza Ubushakashatsi
Ibaruwa y'amakuru y'ubuntu
Ibuka amakuru ya buri munsi ya AI
Inkuru eshatu zizewe za AI buri gitondo cy'iminsi y'akazi, zanditse mu rurimi rworoshye. Ubuntu iteka ryose, nta kwamamaza.
Imeri imwe buri munsi y'akazi. Kwiyimura mu buryo bumwe. Ntitwigera tugurisha cyangwa ngo dusangire aderesi yawe.
Isuzume
Take the Pseudo-Labeling and Self-Training quiz
Ibisubizo ako kanya kuri buri kibazo, n'icyemezo gishobora gusangirwa gifite ID yagenzurwa umaze gutsinda isomo.
Fasha uburezi bwa AI bw'ubuntu. AI Understanding ni umuryango utegamiye kuri leta wa 501(c)(3) — nta kwamamaza, nta kwishyura, iteka ryose. Tanga impano
Ibibazo bikunze kubazwa
What is Pseudo-Labeling and Self-Training?
Pseudo-labeling ni tekinike igenzurwa na kimwe cya kabiri aho icyitegererezo cyatojwe kumurongo muto wanditseho ibyarango byacyo kubirango bitemewe, hanyuma bigatoza kuri ubwo buhanuzi. Nuburyo bworoshye, bukomeye bwo gukoresha amakuru menshi atanditse.
How does a basic self-training loop use unlabeled examples?
The guide describes generating pseudo-labels and training with labeled and selected unlabeled examples.
Why apply a confidence threshold to pseudo-labels?
The guide says high-confidence predictions are retained to reduce contamination from uncertain guesses.
What is confirmation bias in pseudo-labeling?
The guide warns that a model’s errors may be reinforced by its generated labels.
In FixMatch, where does the pseudo-label come from?
FixMatch uses a confident prediction from a weak augmentation to train on a strong augmentation.
What is the student-noise idea in Noisy Student?
The guide describes adding noise during student training to encourage generalization.
Komeza wige
Kuyobora
Abandi bayobozi batoranijwe kuriyi ngingo