{"id":561,"date":"2023-11-25T10:22:20","date_gmt":"2023-11-25T03:22:20","guid":{"rendered":"https:\/\/xuhuongai.com\/?p=561"},"modified":"2023-11-25T19:57:43","modified_gmt":"2023-11-25T12:57:43","slug":"google-deepmind-muon-dinh-ro-khai-niem-ve-tri-tue-nhan-tao-tong-quat","status":"publish","type":"post","link":"https:\/\/xuhuongai.com\/?p=561","title":{"rendered":"Google DeepMind mu\u1ed1n \u0111\u1ecbnh r\u00f5 kh\u00e1i ni\u1ec7m v\u1ec1 tr\u00ed tu\u1ec7 nh\u00e2n t\u1ea1o t\u1ed5ng qu\u00e1t"},"content":{"rendered":"\n
\n

B\u1ea3n tin \u0111\u01b0\u1ee3c t\u00f3m t\u1eaft Anh-Vi\u1ec7t b\u1edfi n\u1ec1n t\u1ea3ng t\u1ea1o tr\u1ee3 l\u00fd AI \u2013 KamiMind<\/a>.<\/p>\nNgu\u1ed3n: Will Douglas Heaven, “Google DeepMind wants to define what counts as artificial general intelligence<\/a>“, MIT Technology Review, 16\/11\/2023.<\/cite><\/blockquote>\n\n\n\n

\"\"
H\u00ecnh minh h\u1ecda. Ngu\u1ed3n: MITTR<\/figcaption><\/figure>\n\n\n\n

Nh\u00f3m nghi\u00ean c\u1ee9u h\u00e0ng \u0111\u1ea7u c\u1ee7a Google DeepMind \u0111\u00e3 \u0111\u01b0a ra m\u1ed9t \u0111\u1ecbnh ngh\u0129a ho\u00e0n to\u00e0n m\u1edbi v\u1ec1 tr\u00ed tu\u1ec7 nh\u00e2n t\u1ea1o t\u1ed5ng qu\u00e1t (AGI). Theo \u0111\u1ecbnh ngh\u0129a m\u1edbi n\u00e0y, kh\u1ea3 n\u0103ng c\u1ee7a AGI kh\u00f4ng ch\u1ec9 d\u1eebng l\u1ea1i \u1edf vi\u1ec7c \u0111\u1ea1t \u0111\u01b0\u1ee3c ho\u1eb7c v\u01b0\u1ee3t qua kh\u1ea3 n\u0103ng c\u1ee7a con ng\u01b0\u1eddi trong nhi\u1ec1u l\u0129nh v\u1ef1c kh\u00e1c nhau. M\u00e0 h\u01a1n th\u1ebf, AGI c\u00f2n ph\u1ea3i c\u00f3 kh\u1ea3 n\u0103ng t\u1ef1 h\u1ecdc \u0111\u1ec3 th\u1ef1c hi\u1ec7n nh\u1eefng nhi\u1ec7m v\u1ee5 \u0111\u00f3 m\u1ed9t c\u00e1ch hi\u1ec7u qu\u1ea3, bi\u1ebft \u0111\u00e1nh gi\u00e1 ch\u1ea5t l\u01b0\u1ee3ng c\u00f4ng vi\u1ec7c c\u1ee7a ch\u00ednh m\u00ecnh v\u00e0 kh\u00f4ng ng\u1ea7n ng\u1ea1i y\u00eau c\u1ea7u s\u1ef1 h\u1ed7 tr\u1ee3 khi c\u1ea7n thi\u1ebft.<\/p>\n\n\n\n

Nh\u00f3m nghi\u00ean c\u1ee9u c\u1ee7a Google DeepMind c\u0169ng \u0111\u00e3 x\u00e2y d\u1ef1ng m\u1ed9t h\u1ec7 th\u1ed1ng ph\u00e2n lo\u1ea1i AGI g\u1ed3m 5 c\u1ea5p \u0111\u1ed9, t\u1eeb “xu\u1ea5t hi\u1ec7n” \u0111\u1ebfn “si\u00eau nh\u00e2n”. C\u1ea5p \u0111\u1ed9 “xu\u1ea5t hi\u1ec7n” ch\u1ec9 \u0111\u1ebfn nh\u1eefng h\u1ec7 th\u1ed1ng AI \u0111\u1ea7u ti\u00ean c\u00f3 kh\u1ea3 n\u0103ng t\u1ef1 h\u1ecdc v\u00e0 th\u1ef1c hi\u1ec7n nhi\u1ec1u t\u00e1c v\u1ee5. C\u1ea5p \u0111\u1ed9 “si\u00eau nh\u00e2n” d\u00e0nh cho nh\u1eefng h\u1ec7 th\u1ed1ng AI c\u00f3 kh\u1ea3 n\u0103ng v\u01b0\u1ee3t qua con ng\u01b0\u1eddi \u1edf h\u1ea7u h\u1ebft m\u1ecdi t\u00e1c v\u1ee5.<\/p>\n\n\n\n

\u0110\u00e1ng ch\u00fa \u00fd, nh\u00f3m nghi\u00ean c\u1ee9u nh\u1ea5n m\u1ea1nh r\u1eb1ng AGI kh\u00f4ng ng\u1ee5 \u00fd s\u1ef1 t\u1ef1 \u0111\u1ed9ng ho\u00e0n to\u00e0n. Tr\u00e1i l\u1ea1i, n\u1ebfu AGI \u0111\u01b0\u1ee3c ph\u00e1t tri\u1ec3n th\u00e0nh c\u00f4ng, kh\u1ea3 n\u0103ng c\u1ee7a n\u00f3 c\u1ea7n \u0111\u01b0\u1ee3c \u0111\u00e1nh gi\u00e1 li\u00ean t\u1ee5c \u0111\u1ec3 \u0111\u1ea3m b\u1ea3o r\u1eb1ng n\u00f3 \u0111ang ho\u1ea1t \u0111\u1ed9ng m\u1ed9t c\u00e1ch an to\u00e0n v\u00e0 hi\u1ec7u qu\u1ea3. \u0110i\u1ec1u n\u00e0y cho th\u1ea5y s\u1ef1 nh\u1eadn th\u1ee9c s\u00e2u s\u1eafc c\u1ee7a nh\u00f3m nghi\u00ean c\u1ee9u v\u1ec1 tr\u00e1ch nhi\u1ec7m x\u00e3 h\u1ed9i c\u1ee7a h\u1ecd trong vi\u1ec7c ph\u00e1t tri\u1ec3n c\u00f4ng ngh\u1ec7 AI.<\/p>\n\n\n\n

V\u1edbi vi\u1ec7c \u0111\u1ecbnh ngh\u0129a m\u1edbi n\u00e0y c\u1ee7a AGI, Google DeepMind \u0111\u00e3 m\u1edf ra m\u1ed9t ph\u01b0\u01a1ng h\u01b0\u1edbng m\u1edbi trong vi\u1ec7c nghi\u00ean c\u1ee9u v\u00e0 ph\u00e1t tri\u1ec3n tr\u00ed tu\u1ec7 nh\u00e2n t\u1ea1o. H\u1ecd kh\u00f4ng ch\u1ec9 \u0111\u1eb7t ra m\u1ee5c ti\u00eau cho AGI ph\u1ea3i \u0111\u1ea1t \u0111\u01b0\u1ee3c kh\u1ea3 n\u0103ng c\u1ee7a con ng\u01b0\u1eddi, m\u00e0 c\u00f2n \u0111\u1eb7t ra y\u00eau c\u1ea7u cao h\u01a1n l\u00e0 AGI ph\u1ea3i bi\u1ebft c\u00e1ch t\u1ef1 h\u1ecdc v\u00e0 c\u1ea3i thi\u1ec7n b\u1ea3n th\u00e2n. \u0110i\u1ec1u n\u00e0y kh\u00f4ng ch\u1ec9 g\u00f3p ph\u1ea7n \u0111\u1ea9y m\u1ea1nh ti\u1ebfn b\u1ed9 c\u1ee7a c\u00f4ng ngh\u1ec7 AI, m\u00e0 c\u00f2n ph\u1ea3n \u00e1nh r\u00f5 r\u00e0ng t\u1ea7m nh\u00ecn xa x\u00f4i v\u00e0 tham v\u1ecdng l\u1edbn lao c\u1ee7a Google DeepMind trong l\u0129nh v\u1ef1c n\u00e0y.<\/p>\n\n\n\n

Ngo\u00e0i ra, vi\u1ec7c Google DeepMind x\u00e2y d\u1ef1ng h\u1ec7 th\u1ed1ng ph\u00e2n lo\u1ea1i AGI c\u0169ng cho th\u1ea5y h\u1ecd \u0111ang n\u1eafm b\u1eaft r\u00f5 r\u00e0ng h\u01b0\u1edbng \u0111i c\u1ee7a c\u00f4ng ngh\u1ec7 AI trong t\u01b0\u01a1ng lai. H\u1ec7 th\u1ed1ng ph\u00e2n lo\u1ea1i n\u00e0y kh\u00f4ng ch\u1ec9 gi\u00fap ng\u01b0\u1eddi d\u00f9ng hi\u1ec3u r\u00f5 h\u01a1n v\u1ec1 kh\u1ea3 n\u0103ng c\u1ee7a AGI, m\u00e0 c\u00f2n gi\u00fap nh\u00f3m nghi\u00ean c\u1ee9u \u0111\u00e1nh gi\u00e1 \u0111\u01b0\u1ee3c ti\u1ebfn tr\u00ecnh c\u1ee7a m\u00ecnh v\u00e0 \u0111\u1eb7t ra nh\u1eefng m\u1ee5c ti\u00eau m\u1edbi cho t\u01b0\u01a1ng lai. C\u00f9ng v\u1edbi \u0111\u00f3, vi\u1ec7c nh\u1ea5n m\u1ea1nh r\u1eb1ng AGI kh\u00f4ng ng\u1ee5 \u00fd s\u1ef1 t\u1ef1 \u0111\u1ed9ng ho\u00e0n to\u00e0n c\u0169ng cho th\u1ea5y s\u1ef1 c\u1ea9n tr\u1ecdng c\u1ee7a Google DeepMind trong vi\u1ec7c \u0111\u1ea3m b\u1ea3o an to\u00e0n v\u00e0 hi\u1ec7u qu\u1ea3 c\u1ee7a AGI.<\/p>\n\n\n\n

B\u1ea3n t\u00f3m t\u1eaft ti\u1ebfng Anh<\/summary>\n

Google DeepMind’s esteemed research team has proposed a novel definition of Artificial General Intelligence (AGI). This fresh perspective suggests that the capabilities of AGI extend beyond simply achieving or surpassing human ability in various domains. Instead, AGI should be capable of autodidacticism, effectively performing tasks independently, assessing its own performance, and seeking assistance when necessary.<\/p>\n\n\n\n

The research team at Google DeepMind has also established a five-tiered classification system for AGI, ranging from “emergent” to “superhuman”. The “emergent” level refers to early AI systems that can self-learn and execute multiple tasks, while the “superhuman” designation is for AI systems that can surpass human capabilities in virtually all tasks.<\/p>\n\n\n\n

Notably, the research team emphasizes that AGI does not imply total autonomy. On the contrary, if AGI is successfully developed, its capabilities will need constant evaluation to ensure it is operating safely and effectively. This demonstrates the team’s profound awareness of their social responsibility in developing AI technology.<\/p>\n\n\n\n

With this new definition of AGI, Google DeepMind is charting a new course in AI research and development. They aim not just for AGI to match human abilities, but for it to be self-improving. This not only propels the advancement of AI technology but also clearly reflects Google DeepMind’s ambitious vision in this field.<\/p>\n\n\n\n

Furthermore, Google DeepMind’s construction of an AGI classification system shows their clear grasp of the future direction of AI technology. This system not only helps users understand the capabilities of AGI but also allows the research team to assess their progress and set new goals for the future. Alongside this, their emphasis that AGI does not imply total autonomy indicates Google DeepMind’s caution in ensuring the safety and effectiveness of AGI.<\/p>\n<\/details>\n","protected":false},"excerpt":{"rendered":"

Nh\u00f3m nghi\u00ean c\u1ee9u h\u00e0ng \u0111\u1ea7u c\u1ee7a Google DeepMind \u0111\u00e3 \u0111\u01b0a ra m\u1ed9t \u0111\u1ecbnh ngh\u0129a ho\u00e0n to\u00e0n m\u1edbi v\u1ec1 tr\u00ed tu\u1ec7 nh\u00e2n t\u1ea1o t\u1ed5ng qu\u00e1t (AGI). Theo \u0111\u1ecbnh ngh\u0129a m\u1edbi n\u00e0y, kh\u1ea3 n\u0103ng c\u1ee7a AGI kh\u00f4ng ch\u1ec9 d\u1eebng l\u1ea1i \u1edf vi\u1ec7c \u0111\u1ea1t \u0111\u01b0\u1ee3c ho\u1eb7c v\u01b0\u1ee3t qua kh\u1ea3 n\u0103ng c\u1ee7a con ng\u01b0\u1eddi trong nhi\u1ec1u l\u0129nh v\u1ef1c kh\u00e1c nhau. M\u00e0 h\u01a1n th\u1ebf, AGI c\u00f2n ph\u1ea3i c\u00f3 kh\u1ea3 n\u0103ng t\u1ef1 h\u1ecdc \u0111\u1ec3 th\u1ef1c hi\u1ec7n nh\u1eefng nhi\u1ec7m v\u1ee5 \u0111\u00f3 m\u1ed9t c\u00e1ch hi\u1ec7u qu\u1ea3, bi\u1ebft \u0111\u00e1nh gi\u00e1 ch\u1ea5t l\u01b0\u1ee3ng c\u00f4ng vi\u1ec7c c\u1ee7a ch\u00ednh m\u00ecnh v\u00e0 kh\u00f4ng ng\u1ea7n ng\u1ea1i y\u00eau c\u1ea7u s\u1ef1 h\u1ed7 tr\u1ee3 khi c\u1ea7n thi\u1ebft.<\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[5],"tags":[84,97,10],"class_list":["post-561","post","type-post","status-publish","format-standard","hentry","category-ai-news","tag-agi","tag-google-deepmind","tag-tri-tue-nhan-tao"],"_links":{"self":[{"href":"https:\/\/xuhuongai.com\/index.php?rest_route=\/wp\/v2\/posts\/561","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/xuhuongai.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/xuhuongai.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/xuhuongai.com\/index.php?rest_route=\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/xuhuongai.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=561"}],"version-history":[{"count":2,"href":"https:\/\/xuhuongai.com\/index.php?rest_route=\/wp\/v2\/posts\/561\/revisions"}],"predecessor-version":[{"id":565,"href":"https:\/\/xuhuongai.com\/index.php?rest_route=\/wp\/v2\/posts\/561\/revisions\/565"}],"wp:attachment":[{"href":"https:\/\/xuhuongai.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=561"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/xuhuongai.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=561"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/xuhuongai.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=561"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}