{"id":4297,"date":"2022-03-31T17:46:19","date_gmt":"2022-03-31T17:46:19","guid":{"rendered":"https:\/\/essentials.pixfort.com\/digital-agency-elementor\/?page_id=4297"},"modified":"2025-06-18T11:40:28","modified_gmt":"2025-06-18T02:40:28","slug":"klue-korean-language-understanding-evaluation","status":"publish","type":"page","link":"https:\/\/open.datumo.com\/en\/?page_id=4297","title":{"rendered":"KLUE – Korean Language Understanding Evaluation Benchmark"},"content":{"rendered":"\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

KLUE <\/span><\/span> <\/h2><\/div><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Korean <\/span><\/span> Language <\/span><\/span> Understanding <\/span><\/span> Evaluation <\/span><\/span> Benchmark <\/span><\/span> <\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\tDownload Datasets<\/span><\/a>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\tLearn about Datumo<\/span><\/a>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t
TAGS<\/h6><\/div><\/div><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t
Korean<\/span><\/span><\/h6>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t
Natural language<\/span><\/span><\/h6>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t
NLP<\/span><\/span><\/h6>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t
Considers unique characteristics of the Korean language<\/h5><\/div><\/div><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Natural language processing(NLP) has been an on-going research throughout the world. However, there were limitations in utilizing open datasets as the basis for Korean NLP research because most of them were in english, which made it difficult to produce precise results taking the unique characteristics of the Korean language into consideration. In order to solve this problem, the startup Upstage held hands with 10 other institutes including Korea Advanced Institute of Science and Technology(KAIST), New York University(NYU), Naver, and Google to build KLUE benchmark, which stands for Korean Language Understanding Evaluation. KLUE aims to build an evaluation system to establish fair comparison across Korean language models. DATUMO has taken part in forming KLUE by providing and processing data.<\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t
\"\"<\/div><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

KLUE is a collection of 8 Korean natural language understanding (NLU) tasks, including Topic Classification, SemanticTextual Similarity, Natural Language Inference, Named Entity Recognition, Relation Extraction, Dependency Parsing, Machine Reading Comprehension, and Dialogue State Tracking. The dataset consists of original Korean used in daily life, rather than Korean translated from English, which allows higher precision in understanding and evaluation of the language. By providing the standards for accurate evaluation of various Korean language models, KLUE is definitely anticipated to accelerate the development of Korean NLP research.<\/strong><\/p>

Also, the significance of the KLUE project lies in the fact that it is the first licensed, open dataset ever in Korea.<\/p><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t
About<\/h6><\/div><\/div><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Korean <\/span><\/span> Language\nUnderstanding <\/span><\/span> Evaluation <\/span><\/span> <\/h3><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t
\"\"<\/div><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Datumo provides high quality data for smarter AI. As part of Datumo's Data Sponsorship Program, Datumo cooperated with Upstage in building the following dataset.<\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t

Upstage was founded by Sunghun Kim, who goes by Sung Kim, on October 2020, in order to accelerate the AI transformation of enterprises. Upstage searches for issues in enterprises that can be improved with AI technology and provides consultation on building basic AI models and systems according to their necessities.<\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t

Sung Kim, the founder of Upstage, has received the Most Influential Paper at ICSME 2018. He is also a four-time winner of the ACM SIGSOFT Distinguished Paper Award (ICSE 2007, ASE 2012, ICSE 2013 and ISSTA 2014). Sung Kim is an Associate Professor of Computer Science at the Hong Kong University of Science and Technology, and has led the Naver Clova AI team in Korea.<\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t

Upstage is planning on supporting enterprises with successful AI transformation by helping them develop their own group of professionals with deeper knowledge in AI. Upstage actively seeks to provide robust education based on their AI business experiences.<\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t
<\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t
Testimonial<\/h5><\/div><\/div><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

\u201cWhile building the KLUE dataset with Datumo, we were most impressed by their data quality\nassurance system. Despite the intricacy of the data and the tight deadline, Datumo was able to provide\nspecific guidelines for the workers to guarantee data consistency. They also made sure to train and select qualified workers, and inspect the entire dataset. We believe that KLUE, the representative Korean NLP benchmark dataset, was able to come into the world, owing to Datumo\u2019s capability and passion.\u201d<\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

\n\t\t\t\t
\n\t\t\t\t\t

SungJoon Park, Upstage AI research engineer & Chief project manager of Project KLUE\n<\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

Dataset <\/span><\/span> specification <\/span><\/span> <\/h3><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t
\n\t\t\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t
TC<\/h5><\/div><\/div><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t

  • 210,000 Classification annotations (70,000 headlines * 3 topics)<\/li><\/ul><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
    \n\t\t\t\t
    \n\t\t\t\t\t
    STS<\/h5><\/div><\/div><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
    \n\t\t\t\t
    \n\t\t\t\t\t

    • 105,000 similarity score labels (15,000 pairs of sentences * 7 similarity scores)<\/li><\/ul><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
      \n\t\t\t\t
      \n\t\t\t\t\t
      NLI<\/h5><\/div><\/div><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
      \n\t\t\t\t
      \n\t\t\t\t\t

      • 30,998 sentence pairs<\/li><\/ul><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
        \n\t\t\t\t
        \n\t\t\t\t\t
        \"\"<\/div><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
        \n\t\t\t\t
        \n\t\t\t\t\t

        Statistics for KLUE-NLI. The first three columns provide the number of sentence pairs for trains, developments, and tests. Set. The mean Len Prem and the mean Len Hyp are the average number of characters in the premise and hypothesis sentences, respectively._Source: KLUE Paper<\/p><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

        \n\t\t\t\t
        \n\t\t\t\t\t
        MRC<\/h5><\/div><\/div><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
        \n\t\t\t\t
        \n\t\t\t\t\t

        • 29,313 questions (Type1: 12,207\/ Type2: 7,895\/ Type3: 9,211)<\/li><\/ul><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t\t
          \n\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t

          Process <\/span><\/span> of <\/span><\/span> annotation <\/span><\/span> <\/h3><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t

          Amongst the eight Korean natural language understanding (NLU) tasks, Datumo took responsibility for four: Topic Classification, Semantic Textual Similarity, Natural Language Inference, and Machine Reading Comprehension.
          Especially for Topic Classification and Semantic Textual Similarity, qualified in-house workers participated for maximum accuracy.<\/p><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

          \n\t\t\t\t
          \n\t\t\t\t\t

          As the host of the \"AI Dataset Sponsorship Program\", Datumo has also participated in the KLUE project as a sponsor.<\/p><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t

          \n\t\t\t\t\t\t
          \n\t\t\t\t\t
          \n\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          Data <\/span><\/span> Collection <\/span><\/span> <\/h6><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t

          Topic <\/span><\/span> Classification <\/span><\/span> <\/h3><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t\t
          \n\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t\t
          \n\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t

          Topic Classification aims to train a classifier to predict the topic of a given text snippet.<\/p>

          1) News headlines from online articles are collected.
          2) For each headline, three annotators independently label three topics in order of relevance among seven categories.
          \u2192 Workers are requested to report any headline that includes personally identifiable information (PII), expresses social bias, or is hate speech.
          The reported headline is discarded after manual review.<\/p><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

          \n\t\t\t\t
          \n\t\t\t\t\t\t\t
          \n\t\t\t
          <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t

          Portion of data were collected and labeled using Cash Mission<\/strong>, Datumo's crowd-sourcing platform.<\/span><\/p><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

          \n\t\t\t\t
          \n\t\t\t\t\t\t\t
          \n\t\t\t
          <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t\t\t\t\t\t
          \n\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t
          Left) Screenshot of project uploaded for data labeling on Cash Mission(M) 'Evaluating senetence similarity' \/ Right) Web screenshot of the guideline for project on Cash Mission 'Classifying News Headline'<\/figcaption>\n\t\t\t\t\t\t\t\t\t\t<\/figure>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t
          \n\t\t\t
          <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t

          Sample Data<\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t

          \n\t\t\t\t
          \n\t\t\t\t\t

          \n{\n    \"guid\": \"ynat-v1_dev_00000\",\n    \"title\": \"5\uc5b5\uc6d0 \ubb34\uc774\uc790 \uc735\uc790\ub294 \ub418\uace0 7\ucc9c\ub9cc\uc6d0 \uc774\uc0ac\ube44\ub294 \uc548\ub41c\ub2e4\",\n    \"predefined_news_category\": \"\uacbd\uc81c\",\n    \"label\": \"\uc0ac\ud68c\",\n    \"annotations\": {\n        \"annotators\": [\n            \"18\",\n            \"03\",\n            \"15\"\n        ],\n        \"annotations\": {\n            \"first-scope\": [\n                \"\uc0ac\ud68c\",\n                \"\uc0ac\ud68c\",\n                \"\uacbd\uc81c\"\n            ],\n            \"second-scope\": [\n                \"\ud574\ub2f9\uc5c6\uc74c\",\n                \"\ud574\ub2f9\uc5c6\uc74c\",\n                \"\uc0ac\ud68c\"\n            ],\n            \"third-scope\": [\n                \"\ud574\ub2f9\uc5c6\uc74c\",\n                \"\ud574\ub2f9\uc5c6\uc74c\",\n                \"\uc0dd\ud65c\ubb38\ud654\"\n            ]\n        }\n    },\n}\n\n<\/pre><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t\t
          \n\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          Data <\/span><\/span> Collection <\/span><\/span> <\/h6><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          Semantic <\/span><\/span> textual <\/span><\/span> similarity <\/span><\/span> <\/h3><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t\t
          \n\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t\t
          \n\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          Semantic textual similarity (STS) is necessary to measure the degree of semantic equivalence between two sentences. STS is
          essential to other NLP tasks such as machine translation, summarization, and question answering.

          1) Compare the semantics of a pair of sentences - Evaluate the similarity based on the purpose or the meaning of the sentence, not
          on the usage of same vocabulary.

          2) Score the similarity by comparing \"important\" and \"unimportant\" contents - \"Important\" contents cover the purpose or the
          meaning, while \"unimportant\" contents cover the nuance or expression of politeness.<\/p><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t
          \n\t\t\t
          <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          Portion of data were collected and labeled using Cash Mission<\/strong>, Datumo's crowd-sourcing platform.<\/span><\/p><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t
          \n\t\t\t
          <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t\t\t\t\t\t
          \n\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t
          Left) Screenshot of project uploaded for data labeling on Cash Mission(M) 'Evaluating sentence similarity' \/ Right) Screenshot of the guideline for project 'Evaluating sentence similarity' on Cash Mission(W)<\/figcaption>\n\t\t\t\t\t\t\t\t\t\t<\/figure>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t
          \n\t\t\t
          <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          Sample Data<\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          \n{\n\u00a0\u00a0\"guid\":\u00a0\"klue-sts-v1_dev_00001\",\n\u00a0\u00a0\"source\":\u00a0\"airbnb-sampled\",\n\u00a0\u00a0\"sentence1\":\u00a0\"\uc8fc\uc694\u00a0\uad00\uad11\uc9c0\u00a0\ubaa8\ub450\u00a0\uac78\uc5b4\uc11c\u00a0\uc774\ub3d9\uac00\ub2a5\ud569\ub2c8\ub2e4.\",\n\u00a0\u00a0\"sentence2\":\u00a0\"\uc704\uce58\ub294\u00a0\ud53c\ub80c\uccb4\u00a0\uc911\uc2ec\uac00\uae4c\uc9c0\u00a0\uac78\uc5b4\uc11c\u00a0\uc774\ub3d9\u00a0\uac00\ub2a5\ud569\ub2c8\ub2e4.\",\n\u00a0\u00a0\"labels\":\u00a0{\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"label\":\u00a01.4,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"real-label\":\u00a01.428571428571429,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"binary-label\":\u00a00\n\u00a0\u00a0},\n\u00a0\u00a0\"annotations\":\u00a0{\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"agreement\":\u00a0\"0:4:3:0:0:0\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"annotators\":\u00a0[\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"16\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"17\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"07\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"03\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"06\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"19\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"14\"\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0],\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"annotations\":\u00a0[\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a02,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a02,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a02,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a01\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0]\n\u00a0\u00a0}\n},\n<\/pre><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t\t
          \n\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          Data <\/span><\/span> Collection <\/span><\/span> <\/h6><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          Natural <\/span><\/span> Language <\/span><\/span> Inference <\/span><\/span> <\/h3><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t\t
          \n\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t\t
          \n\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          The goal of natural language inference (NLI) is to train a model to infer the relationship between the hypothesis sentence and the premise sentence.<\/p>
          1) Based on the premise, annotator makes three hypothesis sentences that correspond to each of the three relationship classes(true\/false\/undetermined).
          2) Four additional annotators label the relationship between the premise and the hypothesis, for validation.<\/p><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t
          \n\t\t\t
          <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t\t\t\t\t\t
          \n\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t
          We call the label intended by the original annotator in writing the hypothesis \"author\u2019s label.\" Consensus among three out of five annotators is \"gold label.\"_Source: KLUE Paper<\/figcaption>\n\t\t\t\t\t\t\t\t\t\t<\/figure>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t
          \n\t\t\t
          <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t
          \n\t\t\t
          <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t
          \n\t\t\t
          <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          KLUE-NLI, based on Datumo's data, performs higher accuracy compared to SNLI(Stanford Natural Language Inference) and MNLI(Multi-Genre Natural Language Inference).<\/p><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t
          \n\t\t\t
          <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          Portion of data were collected and labeled using Cash Mission<\/strong>, Datumo's crowd-sourcing platform.<\/span><\/p><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t
          \n\t\t\t
          <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t\t\t\t\t\t
          \n\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t
           Web screenshot of the tutorial for project 'Evaluating derived sentences'<\/figcaption>\n\t\t\t\t\t\t\t\t\t\t<\/figure>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t
          \n\t\t\t
          <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t\t\t\t\t\t
          \n\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t
          Web screenshot of the guideline for project 'Evaluating derived sentences'<\/figcaption>\n\t\t\t\t\t\t\t\t\t\t<\/figure>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t
          \n\t\t\t
          <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t\t\t\t\t\t
          \n\t\t\t\t\t\t\t\t\t\t\"\"\t\t\t\t\t\t\t\t\t\t\t
           Web screenshot of the guideline for project <Evaluating derived sentences<\/figcaption>\n\t\t\t\t\t\t\t\t\t\t<\/figure>\n\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t\t\t
          \n\t\t\t
          <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          Sample Data<\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          {\n\"guid\": \"klue-nli-v1_dev_00000\",\n\"source\": \"airbnb\",\n\"premise\": \"\ud761\uc5f0\uc790\ubd84\ub4e4\uc740 \ubc1c\ucf54\ub2c8\uac00 \uc788\ub294 \ubc29\uc774\uba74 \ubc1c\ucf54\ub2c8\uc5d0\uc11c \ud761\uc5f0\uc774 \uac00\ub2a5\ud569\ub2c8\ub2e4.\",\n\"hypothesis\": \"\uc5b4\ub5a4 \ubc29\uc5d0\uc11c\ub3c4 \ud761\uc5f0\uc740 \uae08\uc9c0\ub429\ub2c8\ub2e4.\",\n\"gold_label\": \"contradiction\",\n\"author\": \"contradiction\",\n\"label2\": \"contradiction\",\n\"label3\": \"contradiction\",\n\"label4\": \"contradiction\",\n\"label5\": \"contradiction\"\n},<\/pre><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t\t
          \n\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          Data <\/span><\/span> Collection <\/span><\/span> <\/h6><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          Machine <\/span><\/span> Reading <\/span><\/span> Comprehension <\/span><\/span> <\/h3><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t\t
          \n\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t\t
          \n\t\t\t
          \n\t\t\t\t\t\t
          \n\t\t\t\t
          \n\t\t\t\t\t
          Machine reading comprehension (MRC) is a task designed to evaluate models\u2019 abilities to read a given text passage and then answer a question about the passage, that is, its ability of comprehension.<\/p>
          1) Selected workers generate questions and label corresponding answers spans according to the Type.
          2) Three inspectors are assigned for each question type to validate the generated questions and answers.<\/p>
          Question Types<\/p>
          • Question Paraphrasing (Type 1) :
            Type 1 examples focus on paraphrasing the passage sentences when generating questions to reduce word overlap between them. The paraphrasing enables us to validate whether the model can correctly understand the semantics of the paraphrased question and infer the answer.<\/li><\/ul>
            • Multiple-Sentence Reasoning (Type 2) :
              Type 2 examples focus on making questions requiring multiple-sentence reasoning. Multiple-sentence reasoning requires models to derive answers from the questions by reasoning over at least two sentences in the passage.<\/li><\/ul>
              • Unanswerable Questions (Type 3) :
                Type 3 examples are questions unable to be answered within the given passage. We name these as \u2018unanswerable\u2019 questions.<\/li><\/ul><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                \n\t\t\t\t
                \n\t\t\t\t\t\t\t
                \n\t\t\t
                <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                \n\t\t\t\t
                \n\t\t\t\t\t
                Sample Data<\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                \n\t\t\t\t
                \n\t\t\t\t\t
                Questions and answers are formulated based on given text.\n\n<\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                \n\t\t\t\t
                \n\t\t\t\t\t
                \ud55c\uaca8\uc6b8\uc5d0 \ubd04\uaf43 ... \uc9c0\uad6c\ucd0c \uacf3\uacf3 \uc774\uc0c1\uace0\uc628_\ud55c\uacbd_\uad6d\uc81c

                <\/strong><\/h4>
                \uc720\ub7fd\uacfc \ubbf8\uad6d\uc744 \ube44\ub85d\ud574 \uc138\uacc4 \uac01\uad6d\uc5d0 \uc774\uc0c1\uace0\uc628 \ud604\uc0c1\uc774 \ub098\ud0c0\ub098\uace0 \uc788\ub2e4. \uaca8\uc6b8\ucca0 \ud568\ubc15\ub208\uc774 \ub0b4\ub824\uc57c \ud560 \uacf3\uc5d0 \ubd04\uaf43\uae4c\uc9c0 \ud53c\uace0 \uc788\ub2e4. \ub530\ub73b\ud55c \ub0a0\uc528\uc5d0 \ub098\ub4e4\uc774\uac1d\uc774 \ub298\uba74\uc11c \uc74c\uc2dd\uc810\uacfc \uace8\ud504\uc7a5 \ub4f1\uc740 \ubc18\uc0c9\ud558\uace0 \uc788\uc9c0\ub9cc \uaca8\uc6b8\ucca0 \uc637\uc774 \ud314\ub9ac\uc9c0 \uc54a\uc544 \uc758\ub958\uc5c5\uccb4\ub4e4\uc740 \uc6b8\uc0c1\uc774\ub2e4. \uae00\ub85c\ubc8c \uc5d0\ub108\uc9c0 \ud68c\uc0ac\ub4e4\uc740 \uac00\ub729\uc774\ub098 \uc800\uc720\uac00\ub85c \ud798\uaca8\uc6b4 \uc0c1\ud669\uc5d0\uc11c \ub09c\ubc29 \uc218\uc694 \uac10\uc18c\ub77c\ub294 '\uc545\uc7ac'\uae4c\uc9c0 \ub9cc\ub0ac\ub2e4. \ub7ec\uc2dc\uc544 \uae30\uc0c1\ucca0\uc5d0 \ub530\ub974\uba74 \ud639\ub3c5\ud55c \ucd94\uc704\ub85c \uc545\uba85 \ub192\uc740 \ubaa8\uc2a4\ud06c\ubc14\uc758 \ub0ae\uae30\uc628\uc774 22\uc77c(\ud604\uc9c0\uc2dc\uac04) 7\ub3c4<\/strong><\/span>\uae4c\uc9c0 \uce58\uc19f\uc558\ub2e4.<\/span> <\/strong>1936\ub144 \uc774\ud6c4 79\ub144 \ub9cc\uc5d0 \ucd5c\uace0 \uae30\uc628\uc73c\ub85c \uc608\ub144(\ud3c9\uade0 \uc601\ud558 6.5\ub3c4)\ubcf4\ub2e4 10\ub3c4 \uc774\uc0c1 \ub192\ub2e4. \uae30\uc0c1 \uc804\ubb38\uac00\ub4e4\uc740 \uc774\uc0c1\uace0\uc628 \uc6d0\uc778\uc744 \uc5d8\ub2c8\ub1e8 \ud604\uc0c1\uc73c\ub85c \uc124\uba85\ud558\uace0 \uc788\ub2e4. \uc5d8\ub2c8\ub1e8\ub294 \uc801\ub3c4 \ubd80\uadfc \ud574\uc218\uba74 \uc628\ub3c4\uac00 \uc0c1\uc2b9\ud558\ub294 \ud604\uc0c1\uc73c\ub85c \ud3ed\uc6b0\ub098 \uac00\ubb44, \uaca8\uc6b8\ucca0 \uc774\uc0c1\uace0\uc628 \ub4f1\uc758 \uc6d0\uc778\uc774 \ub41c\ub2e4. \uc774\uc0c1\uace0\uc628\uc5d0 \uc5d0\ub108\uc9c0 \ud68c\uc0ac\ub4e4\uc740 \uc804\uc804\uae0d\uae0d\ud558\uace0 \uc788\ub2e4. \ubbf8\uad6d \ub300\ud615 \ud22c\uc790\uc740\ud589\uc778 \uace8\ub4dc\ub9cc\uc0ad\uc2a4\ub294 22\uc77c \ubc1c\ud6a8\ud55c \ubcf4\uace0\uc11c\uc5d0\uc11c \u201c\uc62c\uaca8\uc6b8 \uc774\uc0c1\uace0\uc628\uc73c\ub85c \ub09c\ubc29\uc6a9 \ucc9c\uc5f0\uac00\uc2a4\uc640 \ub4f1\uc720 \uc18c\ube44\uac00 \uac10\uc18c\ud558\uace0 \uc788\ub2e4\u201d\uba70 \u201c\uc6d0\uc720 \uacf5\uae09 \uacfc\uc789\uacfc \uc138\uacc4 \uacbd\uc81c\uc131\uc7a5 \ub454\ud654 \ub4f1\uc5d0 \ub530\ub77c \uc57d\uc138\uac00 \uc774\uc5b4\uc9c0\ub294 \uc5d0\ub108\uc9c0 \uac00\uaca9\uc744 \ub354\uc6b1 \ub04c\uc5b4\ub0b4\ub9b4 \uac83\u201d\uc774\ub77c\uace0 \uc804\ub9dd\ud588\ub2e4. <\/strong><\/span>\uc720\uac00\uac00 \ub354 \ub5a8\uc5b4\uc9c8 \uc218 \uc788\ub2e4\ub294 \uc804\ub9dd\uc774 \uc81c\uae30\ub418\uba74\uc11c \ubbf8\uad6d 3\uc704 \uc5d0\ub108\uc9c0\uc5c5\uccb4 \ucf54\ub178\ucf54\ud544\ub9bd\uc2a4\ub294 \ub7ec\uc2dc\uc544 \uc6d0\uc720\uac1c\ubc1c\uc0ac\uc5c5\uc5d0\uc11c \uc644\uc804\ud788 \ucca0\uc218\ud558\uaca0\ub2e4\uace0 \ubc1c\ud45c\ud588\ub2e4. \uc720\ub7fd \ucd5c\ub300 \uc5d0\ub108\uc9c0\uc5c5\uccb4 \ub85c\uc5f4\ub354\uce58\uc178\uc740 \uc9c0\ub09c 4\uc6d4 \uc778\uc218\ub97c \ubc1c\ud45c\ud55c \uc601\uad6d \ucc9c\uc5f0\uac00\uc2a4\uc5c5\uccb4 BG\uadf8\ub8f9\uc5d0 \ub300\ud574 50\uc5b5\ub2ec\ub7ec \uc0c1\ub2f9\uc758 \ud22c\uc790\ub97c \ucd95\uc18c\ud560 \uacc4\ud68d\uc774\ub2e4.<\/p><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                \n\t\t\t\t
                \n\t\t\t\t\t\t\t
                \n\t\t\t
                <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                \n\t\t\t\t
                \n\t\t\t\t\t
                sentence<\/p>
                • According to The Hydrometeorological Centre of Russia, Moscow, known for its cold weather, warmed up to 7 degrees Celsius in the afternoon of
                  22nd(local time).<\/li><\/ul>
                  • Goldman Sachs, an American investment bank and financial services company, have reported on 22nd that the consumption of natural gas and oil for heating is decreasing and that oversupply of oil and hesitance in global economic growth will even lower the cost per unit of energy.<\/p><\/li><\/ul><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                    \n\t\t\t\t
                    \n\t\t\t\t\t\t\t
                    \n\t\t\t
                    <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                    \n\t\t\t\t
                    \n\t\t\t\t\t
                    Question<\/p>
                    • What was the highest temperature in Moscow on the day Goldman Sachs published their report?<\/li><\/ul><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                      \n\t\t\t\t
                      \n\t\t\t\t\t\t\t
                      \n\t\t\t
                      <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                      \n\t\t\t\t
                      \n\t\t\t\t\t
                      Answer<\/p>
                      • 7 degrees Celsius<\/li><\/ul><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                        \n\t\t\t\t
                        \n\t\t\t\t\t\t\t
                        \n\t\t\t
                        <\/div>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                        \n\t\t\t\t
                        \n\t\t\t\t\t
                        Sample Data<\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                        \n\t\t\t\t
                        \n\t\t\t\t\t
                        {\n\u00a0\u00a0\"annotatable\":\u00a0{\n\u00a0\u00a0\u00a0\u00a0\"parts\":\u00a0[\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"s1s1h1\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"s1s1v1\"\n\u00a0\u00a0\u00a0\u00a0]\n\u00a0\u00a0},\n\u00a0\u00a0\"anncomplete\":\u00a0true,\n\u00a0\u00a0\"sources\":\u00a0[],\n\u00a0\u00a0\"metas\":\u00a0{},\n\u00a0\u00a0\"entities\":\u00a0[\n\u00a0\u00a0\u00a0\u00a0{\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"classId\":\u00a0\"e_1\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"part\":\u00a0\"s1s1v1\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"offsets\":\u00a0[\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0{\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"start\":\u00a0186,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"text\":\u00a0\"\ub7ec\uc2dc\uc544\u00a0\uae30\uc0c1\uccad\uc5d0\u00a0\ub530\ub974\uba74\u00a0\ud639\ub3c5\ud55c\u00a0\ucd94\uc704\ub85c\u00a0\uc545\uba85\u00a0\ub192\uc740\u00a0\ubaa8\uc2a4\ud06c\ubc14\uc758\u00a0\ub0ae\uae30\uc628\uc774\u00a022\uc77c(\ud604\uc9c0\uc2dc\uac04)\u00a07\ub3c4\uae4c\uc9c0\u00a0\uce58\uc19f\uc558\ub2e4.\"\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0}\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0],\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"coordinates\":\u00a0[],\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"confidence\":\u00a0{\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"state\":\u00a0\"pre-added\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"who\":\u00a0[\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"user:ryu\"\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0],\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"prob\":\u00a01\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0},\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"fields\":\u00a0{},\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"normalizations\":\u00a0{}\n\u00a0\u00a0\u00a0\u00a0},\n\u00a0\u00a0\u00a0\u00a0{\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"classId\":\u00a0\"e_2\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"part\":\u00a0\"s1s1v1\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"offsets\":\u00a0[\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0{\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"start\":\u00a0234,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"text\":\u00a0\"7\ub3c4\"\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0}\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0],\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"coordinates\":\u00a0[],\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"confidence\":\u00a0{\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"state\":\u00a0\"pre-added\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"who\":\u00a0[\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"user:ryu\"\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0],\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"prob\":\u00a01\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0},\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"fields\":\u00a0{},\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"normalizations\":\u00a0{}\n\u00a0\u00a0\u00a0\u00a0},\n\u00a0\u00a0\u00a0\u00a0{\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"classId\":\u00a0\"e_1\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"part\":\u00a0\"s1s1v1\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"offsets\":\u00a0[\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0{\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"start\":\u00a0751,\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"text\":\u00a0\"\ubbf8\uad6d\u00a0\ub300\ud615\u00a0\ud22c\uc790\uc740\ud589\uc778\u00a0\uace8\ub4dc\ub9cc\uc0ad\uc2a4\ub294\u00a022\uc77c\u00a0\ubc1c\ud45c\ud55c\u00a0\ubcf4\uace0\uc11c\uc5d0\uc11c\u00a0\"\uc62c\uaca8\uc6b8\u00a0\uc774\uc0c1\uace0\uc628\uc73c\ub85c\u00a0\ub09c\ubc29\uc6a9\u00a0\ucc9c\uc5f0\uac00\uc2a4\uc640\u00a0\ub4f1\uc720\u00a0\uc18c\ube44\uac00\u00a0\uac10\uc18c\ud558\uace0\u00a0\uc788\ub2e4\"\uba70\u00a0\"\uc6d0\uc720\u00a0\uacf5\uae09\u00a0\uacfc\uc789\uacfc\u00a0\uc138\uacc4\u00a0\uacbd\uc81c\uc131\uc7a5\u00a0\ub454\ud654\u00a0\ub4f1\uc5d0\u00a0\ub530\ub77c\u00a0\uc57d\uc138\uac00\u00a0\uc774\uc5b4\uc9c0\ub294\u00a0\uc5d0\ub108\uc9c0\u00a0\uac00\uaca9\uc744\u00a0\ub354\uc6b1\u00a0\ub04c\uc5b4\ub0b4\ub9b4\u00a0\uac83\"\uc774\ub77c\uace0\u00a0\uc804\ub9dd\ud588\ub2e4.\"\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0}\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0],\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"coordinates\":\u00a0[],\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"confidence\":\u00a0{\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"state\":\u00a0\"pre-added\",\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"who\":\u00a0[\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"user:ryu\"\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0],\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"prob\":\u00a01\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0},\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"fields\":\u00a0{},\n\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\"normalizations\":\u00a0{}\n\u00a0\u00a0\u00a0\u00a0}\n\u00a0\u00a0],\n\u00a0\u00a0\"relations\":\u00a0[]\n}<\/pre><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                        \n\t\t\t\t
                        \n\t\t\t\t\t
                        Text annotation results from tagtog<\/p><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
                        \n\t\t\t\t\t\t
                        \n\t\t\t\t\t
                        \n\t\t\t
                        \n\t\t\t\t\t\t
                        \n\t\t\t\t
                        \n\t\t\t\t\t
                        Applications <\/span><\/span> <\/h3><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                        \n\t\t\t\t
                        \n\t\t\t\t\t
                        For those who want to develop<\/p>
                        • high quality Korean NLP services-<\/li>
                        • fluent machine translation, summarization, and question answering services<\/li><\/ul><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                          \n\t\t\t\t
                          \n\t\t\t\t\t\t\t
                          \n\t\t\t\n\t\t\t\t\t\t<\/span>\n\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                          \n\t\t\t\t
                          \n\t\t\t\t\t
                          \"\"<\/div><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t
                          \n\t\t\t\t\t\t
                          \n\t\t\t\t\t
                          \n\t\t\t
                          \n\t\t\t\t\t\t
                          \n\t\t\t\t
                          \n\t\t\t\t\t
                          \"\"<\/div><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t
                          \n\t\t\t
                          \n\t\t\t\t\t\t
                          \n\t\t\t\t
                          \n\t\t\t\t\t
                          CC BY-SA<\/strong><\/p>
                          Reusers are allowed to distribute, remix, adapt, and build upon the material in any medium or format, even commercially, so long as attribution is given to the creator. If you remix, adapt, or build upon the material, you must license the modified material under identical terms.<\/p>
                          https:\/\/creativecommons.org\/licenses\/by-sa\/3.0\/deed.en<\/p><\/p><\/div>\t\t\t\t<\/div>\n\t\t\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t
                          \n\t\t\t\t\t\t
                          \n\t\t\t\t\t
                          \n\t\t\t
                          \n\t\t\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t\t<\/div>\n\t\t<\/div>\n\t\t\t\t\t<\/div>\n\t\t<\/section>\n\t\t\t\t<\/div>\n\t\t","protected":false},"excerpt":{"rendered":"KLUE  Korean  Language  Understanding  Evaluation  Benchmark  Download Datasets Learn about Datumo TAGS Korean Natural language NLP Considers unique characteristics of the Korean language Natural language processing(NLP) has been an on-going research throughout the world. However, there were limitations in utilizing…","protected":false},"author":1,"featured_media":8558,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":{"om_disable_all_campaigns":false,"_monsterinsights_skip_tracking":false,"_monsterinsights_sitenote_active":false,"_monsterinsights_sitenote_note":"","_monsterinsights_sitenote_category":0,"footnotes":""},"class_list":["post-4297","page","type-page","status-publish","has-post-thumbnail","hentry"],"yoast_head":"\nKLUE - Korean Language Understanding Evaluation Benchmark - Datumo Open Datasets for Data-Centric AI<\/title>\n<meta name=\"description\" content=\"KLUE is a Korean natural language comprehension evaluation dataset and is a natural language inference dataset at the level of the CSAT language domain. Datumo provides high-quality learning data for smarter AI.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/open.datumo.com\/en\/?page_id=4297\" \/>\n<meta property=\"og:locale\" content=\"ko_KR\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"KLUE - Korean Language Understanding Evaluation Benchmark - Datumo Open Datasets for Data-Centric AI\" \/>\n<meta property=\"og:description\" content=\"KLUE is a Korean natural language comprehension evaluation dataset and is a natural language inference dataset at the level of the CSAT language domain. Datumo provides high-quality learning data for smarter AI.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/open.datumo.com\/en\/?page_id=4297\" \/>\n<meta property=\"og:site_name\" content=\"Datumo Open Datasets for Data-Centric AI\" \/>\n<meta property=\"article:modified_time\" content=\"2025-06-18T02:40:28+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/open.datumo.com\/en\/wp-content\/uploads\/2023\/06\/kluepreview.png\" \/>\n\t<meta property=\"og:image:width\" content=\"396\" \/>\n\t<meta property=\"og:image:height\" content=\"148\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/png\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\uc608\uc0c1 \ub418\ub294 \ud310\ub3c5 \uc2dc\uac04\" \/>\n\t<meta name=\"twitter:data1\" content=\"11\ubd84\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/?page_id=4297\",\"url\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/?page_id=4297\",\"name\":\"KLUE - Korean Language Understanding Evaluation Benchmark - Datumo Open Datasets for Data-Centric AI\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/?page_id=4297#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/?page_id=4297#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/wp-content\\\/uploads\\\/2023\\\/06\\\/kluepreview.png\",\"datePublished\":\"2022-03-31T17:46:19+00:00\",\"dateModified\":\"2025-06-18T02:40:28+00:00\",\"description\":\"KLUE is a Korean natural language comprehension evaluation dataset and is a natural language inference dataset at the level of the CSAT language domain. Datumo provides high-quality learning data for smarter AI.\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/?page_id=4297#breadcrumb\"},\"inLanguage\":\"ko-KR\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/open.datumo.com\\\/en\\\/?page_id=4297\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"ko-KR\",\"@id\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/?page_id=4297#primaryimage\",\"url\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/wp-content\\\/uploads\\\/2023\\\/06\\\/kluepreview.png\",\"contentUrl\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/wp-content\\\/uploads\\\/2023\\\/06\\\/kluepreview.png\",\"width\":396,\"height\":148},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/?page_id=4297#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"KLUE – Korean Language Understanding Evaluation Benchmark\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/#website\",\"url\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/\",\"name\":\"Datumo Open Datasets for Data-Centric AI\",\"description\":\"ALL IN ONE DATA SOLUTION DATUMO\",\"publisher\":{\"@id\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"ko-KR\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/#organization\",\"name\":\"Datumo Open Datasets for Data-Centric AI\",\"url\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"ko-KR\",\"@id\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/wp-content\\\/uploads\\\/2023\\\/02\\\/datumood_black.png\",\"contentUrl\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/wp-content\\\/uploads\\\/2023\\\/02\\\/datumood_black.png\",\"width\":1835,\"height\":167,\"caption\":\"Datumo Open Datasets for Data-Centric AI\"},\"image\":{\"@id\":\"https:\\\/\\\/open.datumo.com\\\/en\\\/#\\\/schema\\\/logo\\\/image\\\/\"}}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"KLUE - Korean Language Understanding Evaluation Benchmark - Datumo Open Datasets for Data-Centric AI","description":"KLUE is a Korean natural language comprehension evaluation dataset and is a natural language inference dataset at the level of the CSAT language domain. Datumo provides high-quality learning data for smarter AI.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/open.datumo.com\/en\/?page_id=4297","og_locale":"ko_KR","og_type":"article","og_title":"KLUE - Korean Language Understanding Evaluation Benchmark - Datumo Open Datasets for Data-Centric AI","og_description":"KLUE is a Korean natural language comprehension evaluation dataset and is a natural language inference dataset at the level of the CSAT language domain. Datumo provides high-quality learning data for smarter AI.","og_url":"https:\/\/open.datumo.com\/en\/?page_id=4297","og_site_name":"Datumo Open Datasets for Data-Centric AI","article_modified_time":"2025-06-18T02:40:28+00:00","og_image":[{"width":396,"height":148,"url":"https:\/\/open.datumo.com\/en\/wp-content\/uploads\/2023\/06\/kluepreview.png","type":"image\/png"}],"twitter_card":"summary_large_image","twitter_misc":{"\uc608\uc0c1 \ub418\ub294 \ud310\ub3c5 \uc2dc\uac04":"11\ubd84"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/open.datumo.com\/en\/?page_id=4297","url":"https:\/\/open.datumo.com\/en\/?page_id=4297","name":"KLUE - Korean Language Understanding Evaluation Benchmark - Datumo Open Datasets for Data-Centric AI","isPartOf":{"@id":"https:\/\/open.datumo.com\/en\/#website"},"primaryImageOfPage":{"@id":"https:\/\/open.datumo.com\/en\/?page_id=4297#primaryimage"},"image":{"@id":"https:\/\/open.datumo.com\/en\/?page_id=4297#primaryimage"},"thumbnailUrl":"https:\/\/open.datumo.com\/en\/wp-content\/uploads\/2023\/06\/kluepreview.png","datePublished":"2022-03-31T17:46:19+00:00","dateModified":"2025-06-18T02:40:28+00:00","description":"KLUE is a Korean natural language comprehension evaluation dataset and is a natural language inference dataset at the level of the CSAT language domain. Datumo provides high-quality learning data for smarter AI.","breadcrumb":{"@id":"https:\/\/open.datumo.com\/en\/?page_id=4297#breadcrumb"},"inLanguage":"ko-KR","potentialAction":[{"@type":"ReadAction","target":["https:\/\/open.datumo.com\/en\/?page_id=4297"]}]},{"@type":"ImageObject","inLanguage":"ko-KR","@id":"https:\/\/open.datumo.com\/en\/?page_id=4297#primaryimage","url":"https:\/\/open.datumo.com\/en\/wp-content\/uploads\/2023\/06\/kluepreview.png","contentUrl":"https:\/\/open.datumo.com\/en\/wp-content\/uploads\/2023\/06\/kluepreview.png","width":396,"height":148},{"@type":"BreadcrumbList","@id":"https:\/\/open.datumo.com\/en\/?page_id=4297#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/open.datumo.com\/en\/"},{"@type":"ListItem","position":2,"name":"KLUE – Korean Language Understanding Evaluation Benchmark"}]},{"@type":"WebSite","@id":"https:\/\/open.datumo.com\/en\/#website","url":"https:\/\/open.datumo.com\/en\/","name":"Datumo Open Datasets for Data-Centric AI","description":"ALL IN ONE DATA SOLUTION DATUMO","publisher":{"@id":"https:\/\/open.datumo.com\/en\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/open.datumo.com\/en\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"ko-KR"},{"@type":"Organization","@id":"https:\/\/open.datumo.com\/en\/#organization","name":"Datumo Open Datasets for Data-Centric AI","url":"https:\/\/open.datumo.com\/en\/","logo":{"@type":"ImageObject","inLanguage":"ko-KR","@id":"https:\/\/open.datumo.com\/en\/#\/schema\/logo\/image\/","url":"https:\/\/open.datumo.com\/en\/wp-content\/uploads\/2023\/02\/datumood_black.png","contentUrl":"https:\/\/open.datumo.com\/en\/wp-content\/uploads\/2023\/02\/datumood_black.png","width":1835,"height":167,"caption":"Datumo Open Datasets for Data-Centric AI"},"image":{"@id":"https:\/\/open.datumo.com\/en\/#\/schema\/logo\/image\/"}}]}},"_links":{"self":[{"href":"https:\/\/open.datumo.com\/en\/index.php?rest_route=\/wp\/v2\/pages\/4297","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/open.datumo.com\/en\/index.php?rest_route=\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/open.datumo.com\/en\/index.php?rest_route=\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/open.datumo.com\/en\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/open.datumo.com\/en\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=4297"}],"version-history":[{"count":417,"href":"https:\/\/open.datumo.com\/en\/index.php?rest_route=\/wp\/v2\/pages\/4297\/revisions"}],"predecessor-version":[{"id":8852,"href":"https:\/\/open.datumo.com\/en\/index.php?rest_route=\/wp\/v2\/pages\/4297\/revisions\/8852"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/open.datumo.com\/en\/index.php?rest_route=\/wp\/v2\/media\/8558"}],"wp:attachment":[{"href":"https:\/\/open.datumo.com\/en\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=4297"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}