Prevod hovoren\u00e9ho slova na p\u00edsan\u00fd text dnes u\u017e nie je tak atrakt\u00edvnou t\u00e9mou ako v minulosti. Ka\u017ed\u00fd z n\u00e1s m\u00e1 vo vlastnom mobile aplik\u00e1cie, ktor\u00e9 dok\u00e1\u017eu konvertova\u0165 re\u010d na text aj v sloven\u010dine a s mimoriadnou presnos\u0165ou. Nev\u00fdhody s\u00fa\u010dasn\u00fdch syst\u00e9mov spo\u010d\u00edvaj\u00fa v tom, \u017ee nevedia spr\u00e1vne odhadn\u00fa\u0165, kedy nap\u00edsa\u0165 \u010diarku, bodku alebo poml\u010dku.\u00a0 Aj to je vecou nieko\u013ek\u00fdch rokov.<\/p>\n
Potenci\u00e1l technol\u00f3gi\u00ed konvertuj\u00facich re\u010d na text je ale e\u0161te ove\u013ea v\u00e4\u010d\u0161\u00ed a najmodernej\u0161ie vedeck\u00e9 projekty dokazuj\u00fa, \u017ee sa d\u00e1 s vysokou presnos\u0165ou na text prepisova\u0165 aj slovo, ktor\u00e9 e\u0161te ani nevyslovili. Sta\u010d\u00ed, aby ste na neho mysleli.<\/p>\n
Technol\u00f3gie naz\u00fdvan\u00e9 brain-computer interface (BCI)<\/strong> dok\u00e1\u017eu zabezpe\u010dova\u0165 komunik\u00e1ciu medzi po\u010d\u00edta\u010dom a mozgom a potenci\u00e1l ich vyu\u017eite\u013enosti je priam mimoriadny. Postupne n\u00e1m odha\u013euj\u00fa svet, v ktorom bud\u00fa po\u010d\u00edta\u010de skuto\u010dne reagova\u0165 na podnety od n\u00e1s, ktor\u00e9 sme im nezadali fyzicky, \u010di u\u017e mechanicky alebo hlasom.<\/p>\n V\u00fdznamn\u00fd krok na poli v\u00fdvoja t\u00fdchto technol\u00f3gi\u00ed najnov\u0161ie ohl\u00e1sila vedeck\u00e1 skupina Chang Lab z University of California San Francisco (Joseph G. Makin, David A. Moses & Edward F. Chang), ktor\u00e9 v\u00fdsledky svojej pr\u00e1ce publikovala v magaz\u00edne Nature a k\u00f3d svojho programu publikovala na GitHube. Podstatou pr\u00e1ce tejto vedeckej skupiny je prekladanie takzvan\u00e9ho elektrokortikogram<\/em>, teda elektromagnetick\u00fdch sign\u00e1lov v mozgu, ktor\u00e9 s\u00fa pova\u017eovan\u00e9 za prejavy my\u0161lienok, respekt\u00edve vypovedan\u00fdch i nevypovedan\u00fdch slov.<\/p>\n K priazniv\u00fdm v\u00fdsledkom testov pritom t\u00edm dospel sk\u00f4r n\u00e1hodou, ke\u010f v r\u00e1mci in\u00e9ho projektu sledoval \u010dinnos\u0165 mozgu pacientov s epilepsiou.<\/p>\n Program vyu\u017e\u00edva syst\u00e9my strojov\u00e9ho u\u010denia – alebo umelej inteligencie. Testovan\u00e9 osoby dostali za \u00falohu vypoveda\u0165 30 a\u017e 50 viet, ktor\u00e9 syst\u00e9m tejto umelej inteligencie zaznamenal, porovnal s \u00fadajmi z elektrokortikogramu a s\u00e1m vyhodnotil, ako ho spr\u00e1vne \u010d\u00edta\u0165. N\u00e1sledne u\u017e syst\u00e9m bol schopn\u00fd s\u00e1m rozl\u00ed\u0161i\u0165 \u010fal\u0161ie elektromagnetick\u00e9 sign\u00e1ly a prepisova\u0165 ich na p\u00edsan\u00fd text.<\/p>\n Tak\u00e9to technol\u00f3gie u\u017e dnes ale vedeck\u00ed pracovn\u00edci k dispoz\u00edcii maj\u00fa a vedia s nimi dos\u0165 dobre pracova\u0165. \u010co je \u00faplnou novinkou, je ve\u013emi n\u00edzka miera nespr\u00e1vnych odhadov. U jednej z testovan\u00fdch os\u00f4b sa prejavila chyba iba v 3% konverzi\u00ed my\u0161lienok na text, \u010do je historicky nev\u00eddan\u00fd \u00fadaj.<\/p>\n Podstatou syst\u00e9mov umelej inteligencie je ale schopnos\u0165 u\u010di\u0165 sa vlastn\u00fdmi sk\u00fasenos\u0165ami, teda je rozumn\u00e9 predpoklada\u0165, \u017ee v\u00fdsledky bud\u00fa \u010doskoro e\u0161te lep\u0161ie a presnej\u0161ie.<\/p>\n Zmysel takejto technol\u00f3gie spo\u010d\u00edva nielen v zjednodu\u0161en\u00ed ovl\u00e1dania po\u010d\u00edta\u010da, ale aj pomoci \u013eu\u010fom, ktor\u00ed zo zdravotn\u00fdch d\u00f4vod nie s\u00fa schopn\u00ed rozpr\u00e1va\u0165. Prevod nevypovedan\u00fdch slov im m\u00f4\u017ee v zna\u010dnej miere pom\u00f4c\u0165 komunikova\u0165 s okol\u00edm.<\/p>","protected":false},"excerpt":{"rendered":" Na hlasov\u00fdch asistentov od Google a Amazonu si pomaly zvyk\u00e1me, ale technol\u00f3gie mo\u017eno pos\u00fava\u0165 e\u0161te v\u00fdrazne \u010falej. Slov\u00e1 nemus\u00edme ani len vyslovi\u0165, aby im po\u010d\u00edta\u010d rozumel.<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[51],"tags":[],"class_list":["post-6469","post","type-post","status-publish","format-standard","hentry","category-archiv","two-columns"],"_links":{"self":[{"href":"https:\/\/instrid.sk\/en\/wp-json\/wp\/v2\/posts\/6469","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/instrid.sk\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/instrid.sk\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/instrid.sk\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/instrid.sk\/en\/wp-json\/wp\/v2\/comments?post=6469"}],"version-history":[{"count":0,"href":"https:\/\/instrid.sk\/en\/wp-json\/wp\/v2\/posts\/6469\/revisions"}],"wp:attachment":[{"href":"https:\/\/instrid.sk\/en\/wp-json\/wp\/v2\/media?parent=6469"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/instrid.sk\/en\/wp-json\/wp\/v2\/categories?post=6469"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/instrid.sk\/en\/wp-json\/wp\/v2\/tags?post=6469"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}