Azịza ajụjụ anya
Azịza ajụjụ a na-ahụ anya (VQA) na-ahapụ sistemụ ịza ajụjụ asụsụ okike n'efu gbasara onyonyo, dịka 'mmadụ ole na-eyi okpu?' Ọ na-achọ ijikọ ọnụ ịghọta ma foto ahụ na ajụjụ ahụ iji wepụta azịza ziri ezi.
Ime miri emi
Azịza ajụjụ a na-ahụ anya na-ejikọta ọhụụ kọmputa na nhazi asụsụ okike: nyere onyonyo na ajụjụ, ihe nlereanya ahụ weghachiri azịza, nke nwere ike ịbụ otu okwu, nkebiokwu dị mkpirikpi, ma ọ bụ nzaghachi ee / mba. VQA dataset (Antol et al., 2015) na ụdị VQA v2.0 a nụchara anụcha na-ewuli ọrụ ahụ, nke azịza ziri ezi na-akụda ụdị n'ichepụta site na ederede naanị. Sistemu na-etinye ihe onyonyo na ajụjụ ahụ, jikọta ihe nnọchi anya abụọ ahụ, wee buru amụma azịza, n'akụkọ ihe mere eme site n'ịkewa n'usoro okwu azịza edobere. Taa, nnukwu ụdị asụsụ dị ka GPT-4V, LLaVA, na PaLI na-ejikwa VQA mepere emepe, na-atụgharị uche gbasara ihe, njirimara, ọnụ ọgụgụ, mmekọrịta gbasara ohere, na ọbụna ederede edere n'ime onyonyo.
Nghọta nka nka
Ụdị VQA na-ahụkarị na-etinye ihe oyiyi ahụ (CNN ma ọ bụ ihe ntụgharị ọhụụ) na ajụjụ (transformer text encoder), wee na-ejikọta ha, na-ejikarị nlebara anya ka okwu ajụjụ na-aga na mpaghara oyiyi. Vector jikọtara ọnụ na-enye nhazi ọkwa maka azịza nkịtị ma ọ bụ ihe ngbanwe asụsụ maka nzaghachi mepere emepe. Ọnyà amaara bụ nhụsianya asụsụ: ụdị nwere ike irigbu ọnụ ọgụgụ azịza wee leghara onyonyo a anya, nke na-echekwa datasets dị ka VQA v2.0 kpọmkwem.
Mmetụta atụmatụ
Ọsọ na ọnụ ọgụgụ
Visual AI nwere ike megharịa nyocha, nchọpụta na mkpado ọrụ n'ọtụtụ.
Mee nhọrọ
Otu ndị na-emepụta ihe nwere ike imepụta echiche ngwa ngwa site na ngbanwe akwụkwọ ntuziaka ole na ole.
Team na usoro ọrụ
Ọrụ nwere ike iji onyonyo na akara vidiyo siri ike ịhazi.
Ọdịnihu nke azịza ajụjụ anya
VQA na-etolite site na nhazi nkenke azịza gaa na nhụsịrị anya mepere emepe, ọtụtụ nzọụkwụ nwere nkọwa. Na-atụ anya njikwa ngụkọ, eserese, eserese, na ederede n'ime onyonyo (akwụkwọ VQA), gbakwunyere VQA vidiyo na-eme ka oge na-aga. Mbelata mkparị ụzọ mkpirisi na nhụta anya ka bụ ihe kacha mkpa, dịkwa ka azịza na-agbadata na mpaghara onyonyo a kapịrị ọnụ maka ntụkwasị obi. Ndị enyemaka multimodal nwere ikike ga-azawanye ajụjụ a na-ahụ anya na mkparịta ụka na ekwentị, na robotics, yana n'ime ngwa nnweta nke na-enyere ndị ọrụ aka ịjụ gburugburu ha ajụjụ.
Mmejuputa n'ezie n'ụwa
Ikwe ka ndị kpuru ìsì see ngwaahịa wee jụọ 'Olee ụtọ nke a?' ma ọ bụ 'Gịnị bụ ngafe ụbọchị?'
Na-aza ajụjụ gbasara eserese, ụdị, na dọkụmentị nyochara (akwụkwọ VQA) na usoro ọrụ azụmahịa.
Ndị na-enyere aka ịzụ ahịa ahịa na e-azụmahịa na-aza 'Jaket a ọ nwere mkpuchi?' site na foto ngwaahịa
Na-akwado nyocha ahụike ma ọ bụ ihe onyonyo sayensị site na ịza ajụjụ ezubere iche gbasara nyocha ma ọ bụ onyonyo microscope
Ihe ize ndụ & okporo ụzọ nche
Ikike onyonyo na nkwenye nwere ike bụrụ ihe egwu dị n'iwu ma ọ bụrụ na edoghị anya.
Ọrụ nlereanya nwere ike ịdịgasị iche n'ofe ọkụ, igwe mmadụ, na gburugburu.
Enwere ike ghara ịhụ ihe dị mma ma ọ bụrụ na enyochaghị oke ntụkwasị obi.
Map mmejuputa
Kọwaa ụkpụrụ nnabata maka nkenke, icheta, na ụgwọ njehie.
Nwalee na data dabara na ọnọdụ mmepụta n'ezie.
Tinye nyocha mmadụ maka obere obi ike ma ọ bụ amụma mmetụta dị elu.
Sochie ihe nlere anya wee megharịa ka emechara mgbanwe igwefoto ma ọ bụ dataset.
Nọgide na-eme nchọpụta
Free newsletter
Get the daily AI briefing
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the Visual Question Answering quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
Ntuziaka na-esote
Azịza ajụjụ
Ajụjụ a na-ajụkarị
What is Visual Question Answering?
Azịza ajụjụ a na-ahụ anya (VQA) na-ahapụ sistemụ ịza ajụjụ asụsụ okike n'efu gbasara onyonyo, dịka 'mmadụ ole na-eyi okpu?' Ọ na-achọ ijikọ ọnụ ịghọta ma foto ahụ na ajụjụ ahụ iji wepụta azịza ziri ezi.
Kedu ntinye abụọ nke sistemu azịza ajụjụ anya na-ewe?
VQA na-ewe ma onyonyo yana ajụjụ gbasara ya, wee wepụta azịza dabere na onyonyo a.
Kedu ihe kpatara ejiri jiri azịza 'kwesịrị ekwesị' mepụta VQA v2.0 dataset?
VQA v2.0 jikọtara ajụjụ abụọ na onyonyo nwere azịza dị iche iche, na-amanye ụdị ka ha jiri ntinye anya n'ezie kama iji ọnụ ọgụgụ ederede eme ihe.
Kedu ihe bụ 'ịhụnanya asụsụ' na VQA?
Ọgbaghara asụsụ bụ mgbe ihe nlere na-erigbu ọnụ ọgụgụ azịza jikọtara ya na nkebiokwu ajụjụ kama ilele onyonyo a.
Kedu ka ọtụtụ ụdị VQA si ejikọta ozi na ajụjụ?
Ụdị VQA na-edobe usoro nke ọ bụla wee na-ejikọta ha, na-ejikarị nlebara anya n'ofefe wee jụọ ajụjụ na-aga na mpaghara oyiyi dị mkpa.
Sistemu VQA kpochapụwo na-emepụtakarị azịza site n'ime gịnị?
Ọtụtụ ụdị VQA nke mbụ lebara ọrụ ahụ anya dị ka nkewa maka azịza ndị na-adịkarị, ọ bụ ezie na ụdị ọgbara ọhụrụ na-ewepụta ederede mepere emepe.