There is Wikipedia 'the bot made' with the second largest number of articles in the world



Wikimedia Foundation is that the world's largest online encyclopedia 'that operates

Wikipedia ' everyone can participate in the editing freely, we are multi-language deployment by translating the English version to be the main. However, in recent years, it has been pointed out that the reliability of bots using machine translation has been impaired due to multilingualization. Kyle Wilson, the maintainer of the English version of Wikipedia, discusses the problem of bots over Wikipedia, using 'Wikipedia, the second largest number of articles in the world due to bot translation' as an example.

The World's Second Largest Wikipedia Is Written Almost Entirely by One Bot-VICE
https://www.vice.com/en_us/article/4agamm/the-worlds-second-largest-wikipedia-is-written-almost-entirely-by-one-bot



Wikipedia has stated that one of its goals is 'to make it possible for anyone to freely gain knowledge online, and to make it available in as many languages as possible'. As of 2020, Wikipedia's main language is English and has over 6 million articles, but each page of Wikipedia has been translated into various languages to achieve its goals.

The second largest article in the world is the Cebuano version of Wikipedia, which has about 5.37 million articles, 630,000 less than the English version. Next to the Cebuano version are the German version with 2.98 million articles and the Swedish version with 1.64 million articles.

Many people are more unfamiliar with 'Cebuano' than Swedish or German. Cebuano is a language spoken in the Philippines, such as Cebu, and is said to have about 16.5 million speakers. The Cebuano version of Wikipedia supports 5.37 million articles with only 6 administrators and 14 active users, whereas the English version has 1143 administrators and 137,368 active users Exists.

Why is the Cebuano version so many articles supported? Solving the mystery is the existence of bots.

Wikipedia has a page called `` Cebuano Wikipedia '', which says, `` It boasts more than 5 million articles and is the second largest Wikipedia after the English version, but many of the articles are by Internet bot account Lsjbot It was created. '

Cebuano Wikipedia-Wikipedia



A survey published in November 2019 indicated that there were a total of 1601 bots in each language version of Wikipedia. In the English version and some language versions, there are places where bots are used for simple work and repetitive work, but there are also places where bots are used for writing articles themselves.

Kyle points out that when a single bot is writing many articles on Wikipedia in one language, its quality will be negatively affected. For example, the Cebuano version of Wikipedia uses ' Lsjbot ', which was created by Swedish physicist Swarkel Johansson.

It is known that Lsjbot was involved in 24 million edits out of 29.5 million edits made in the Cebuano language version, and a survey of Wikipedia administrator Guilherme Morandini showed that 99.12 of the creation of Cebuano version articles. It was shown that Lsjbot was involved in%. The survey also shows that all but the top 35 editors are bots, with the exception that there are no humans in the top ten. In response, Morandini stated that 'bots are taking over the human editor.'

Some administrators claim that the bot Lsjbot is `` an existence that hinders article creation, '' while an administrator named Railay Huntley conducted a survey and randomly extracted and analyzed 1000 articles created by Lsjbot However, he stated that 'many are very well organized'.



'The Foundation is trying to solve this problem by providing tools, resources and partnerships to the local language community,' said Adila Svitak of the Wikimedia Foundation. These resources are like

Wikimedia cloud services that allow you to create bots and tools. Svitak says, 'We leave the bot-related policy to each community,' referring to technologies that reduce the burden on editors, such as content translation . I have not denied the existence of bots.

Some of these guidelines emphasize quality, while others focus on numbers that try to write as many articles as possible, even if the content is inadequate. And in the latter case, articles using machine translation left problems. For example, in May 2019, the problem that `` village pump (village pump) '' was written as `` bomba do Village (bombing the village) '' in Portuguese was discussed as a result of using Google machine translation on Wikipedia Was . As a result of posting many such translated articles, Wikipedia itself may be reduced in reliability, and the translation guideline's `` Machine translation '' item says `` Please do not post machine translation as it is '' You.

Lsjbot as not the best way to automate the creation of Wikipedia articles in different languages, Mr. Alexander Pechan skiing and Mr. Erica Azzurini in 2018, the content transformer crew John bot that depends on more human creation I did . This tool, called 'Mbabel', creates drafts based on information in an online database called ' Wikidata ' owned by the Wikimedia Foundation. Unlike the Wikimedia Foundation's content translation tool, Mbabel does not post articles directly, but instead creates a 'user test page on Wikipedia'. It assumes that humans extend the page based on the base created by Mbabel.



The disadvantage of this tool, however, was that its contents depended on Wikidata information. To create a high quality draft, the available Wikidata information must be of high quality. After conducting a demonstration experiment, Azzurini said, 'Of course, it's up to each community to decide what to do with bot-written content. However, in my opinion, treating this type of text equivalent to an encyclopedia is Wikipedia. Wikipedia articles related to other auto-generated content, or Wikipedia itself, could be distrusted. ' Ms. Azzurini emphasized that Mbabel should not be used as a bot, but rather as a tool that assumes that humans edit it for publication.

Early English Wikipedia is also full of articles created by bots, similar to Cebuano, and until 2006 there was a `` bot approval group '' and supervised approved bots, but then English Wikipedia became Created a section policy that prohibits the use of bots in content creation.

Creating a multilingual version of Wikipedia with a single bot is a double-edged sword. It must be better than nothing, but it can lead to quality concerns. Wilson says that while involving the bot, it will eventually need to be evaluated and improved by editors who are multilingual. One day, the bot may be able to do human-level work, but until then, bots like the Cebuano version of Wikipedia were considered `` pilot version Wikipedia '', humans modified It is thought that we should go.

in Note, Posted by darkhorse_log