{"id":293517,"date":"2026-01-20T07:00:10","date_gmt":"2026-01-20T07:00:10","guid":{"rendered":"https:\/\/www.europesays.com\/ie\/293517\/"},"modified":"2026-01-20T07:00:10","modified_gmt":"2026-01-20T07:00:10","slug":"deepseek-technique-to-improve-ais-ability-to-read-long-texts-questioned-by-new-research","status":"publish","type":"post","link":"https:\/\/www.europesays.com\/ie\/293517\/","title":{"rendered":"DeepSeek technique to improve AI\u2019s ability to \u2018read\u2019 long texts questioned by new research"},"content":{"rendered":"<p datatype=\"p\" data-qa=\"Component-Component\" class=\"e8zc9q40 css-1c6uqr6 ec74h0k1\">A group of researchers from China and Japan has challenged a method unveiled several months ago by Chinese artificial intelligence start-up DeepSeek that was designed to improve AI\u2019s ability to handle long blocks of text, marking a rare case of the company\u2019s research being publicly questioned.<\/p>\n<p datatype=\"p\" data-qa=\"Component-Component\" class=\"e8zc9q40 css-1c6uqr6 ec74h0k1\">The DeepSeek-OCR (optical character recognition) method, designed to compress text by using visual representations, potentially revolutionising how AI models handle long texts, was flawed due to inconsistent performance, according to researchers from Japan\u2019s Tohoku University and the Chinese Academy of Sciences.<\/p>\n<p datatype=\"p\" data-qa=\"Component-Component\" class=\"e8zc9q40 css-1c6uqr6 ec74h0k1\">In their study, titled \u201cVisual Merit or Linguistic Crutch? A Close Look at DeepSeek-OCR\u201d, the research team found that the start-up\u2019s method relied heavily on language priors \u2013 the tendency of AI models to draw on patterns learned from large volumes of text \u2013 rather than the visual understanding it claimed, making performance metrics reported by the Chinese company \u201cmisleading\u201d.<\/p>\n<p datatype=\"p\" data-qa=\"Component-Component\" class=\"e8zc9q40 css-1c6uqr6 ec74h0k1\">AI models faced a critical limitation known as the long-context bottleneck, which restricted their ability to process lengthy documents or extended conversations, the researchers noted.<\/p>\n<p datatype=\"p\" data-qa=\"Component-Component\" class=\"e8zc9q40 css-1c6uqr6 ec74h0k1\">Improvements in this area, which would lead to a performance leap in an AI system, have been sought by companies and research institutes worldwide.<\/p>\n<p><img decoding=\"async\" alt=\"A DeepSeek display at an AI fair in Hangzhou, east China\u2019s Zhejiang Province, May 4, 2025. Photo: Xinhua\" data-qa=\"BaseImage-handleRenderImage-StyledImage\" class=\"e1gf69pb2 css-6ikqhs e445x7d0\" loading=\"lazy\" src=\"https:\/\/www.europesays.com\/ie\/wp-content\/uploads\/2026\/01\/a2137d2d-6855-4d82-ba0d-d13b02d5df3b_d2851519.jpg\" title=\"A DeepSeek display at an AI fair in Hangzhou, east China\u2019s Zhejiang Province, May 4, 2025. Photo: Xinhua\"\/>A DeepSeek display at an AI fair in Hangzhou, east China\u2019s Zhejiang Province, May 4, 2025. Photo: Xinhua<\/p>\n<p datatype=\"p\" data-qa=\"Component-Component\" class=\"e8zc9q40 css-1c6uqr6 ec74h0k1\">The DeepSeek-OCR technique, published in October, was said to be able to handle large and complex documents by using visual perception as a compression medium.<\/p>\n","protected":false},"excerpt":{"rendered":"A group of researchers from China and Japan has challenged a method unveiled several months ago by Chinese&hellip;\n","protected":false},"author":2,"featured_media":293518,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[261],"tags":[291,289,290,1793,381,4563,8736,146148,18,19,17,386,146149,82,48448,13861,146150],"class_list":{"0":"post-293517","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-artificial-intelligence","8":"tag-ai","9":"tag-artificial-intelligence","10":"tag-artificialintelligence","11":"tag-beijing","12":"tag-china","13":"tag-chinese-academy-of-sciences","14":"tag-deepseek","15":"tag-deepseek-ocr","16":"tag-eire","17":"tag-ie","18":"tag-ireland","19":"tag-japan","20":"tag-li-bojie","21":"tag-technology","22":"tag-tohoku-university","23":"tag-university-of-science-and-technology-of-china","24":"tag-visual-question-answering"},"share_on_mastodon":{"url":"https:\/\/pubeurope.com\/@ie\/115926140372020889","error":""},"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/posts\/293517","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/comments?post=293517"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/posts\/293517\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/media\/293518"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/media?parent=293517"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/categories?post=293517"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ie\/wp-json\/wp\/v2\/tags?post=293517"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}