YOU SAID:

Text analytics allows us to work with knowledge that has been stored as written language. Humans have used written language to store and transmit knowledge for thousands of years. Advances in the production of written documents have fueled major social and scientific revolutions, from the invention of writing itself to the printing press to regular postal networks to the internet. Each of these revolutionary advances has made it easier for more people to create, share, and, therefore to understand more types of written language. With each advance, our societies have produced exponentially more texts. The internet and the production of digital language data have again revolutionized the flow of information. Newspapers are almost entirely online, most people access articles through social media, and informal (micro)blogs have dislodged experts as primary sources of information. These changes have dramatically increased the amount of written language that we need to work with. In the past, those who couldn’t read or write faced major social and economic obstacles. In the future, those who are technologically illiterate will face similar obstacles. We now have millions of news articles, billions of Tweets, trillions of web pages to understand and process. So, how do we make sense of this vast amount of information contained in written language? The answer is artificial intelligence (AI). Before the internet, the problem was how to find specific knowledge and information. But now we have too much knowledge to choose from. AI allows us to automate the searching and classifying process so that we can focus on reading things that actually matter to us. In the past getting multiple views on a subject was a challenge. But now the challenge is to understand the views of millions of writers who are each talking to a different audience. The information available is abundant but fragmented. AI therefore allows us to analyze and synthesize information in a way that facilitates individual comprehension. From an ethical perspective, in the past the problem was to understand people whose views had been left outside and marginalized: minorities, indigenous peoples, and women. But now the issue is how to make sure that languages and dialects are captured and represented through digital technologies: AI allows us to adapt language analysis models based on English to other languages. Before the Internet Now Information Access Hard to find specific information Too much information to choose from: hard to filter through what is available Point of View Hard to get multiple views on a subject Millions of views on every subject: hard to understand a fragmented population Representation Hard to represent the views of marginalized populations Hard to ensure that minority languages and dialects are represented digitally But how does AI compare with human abilities for such tasks? What does language mean to a machine? · Sometimes AI reproduces human intuitions about the meaning of a document. For example, we present a case-study on socio-economic indicators in which we use text analytics to determine whether a news article is talking about corruption. Here, the output of the text classifier is similar to a human annotating the subject of an article. · In other cases AI detects hidden meaning in documents that humans are unable to see. For example, we have another case-study in digital humanities where we use text analytics to model authorship in books. Although we as humans are usually unable to determine who wrote what, AI can easily find enough clues to accurately predict authorship. · In other cases AI completely misses something that is easily accessible to humans. For example, we will look at a case-study on multi-lingualism and the internet where we see that 90% of language data represents just 20 languages. Minority languages have very little data available. This means that AI is unable to learn about these languages and thus unable to access the cultural knowledge they contain. Just like individual humans, AI can’t work with languages it hasn’t learned to read.

INTO JAPANESE

テキスト分析により、書き言葉として保存されている知識を使用することができます。 人間は何千年もの間、知識を保存し、伝達するために書かれた言語を使用してきました。 文書の作成の進歩は、執筆自体の発明から印刷機、再

BACK INTO ENGLISH

Text analysis allows you to use knowledge that is stored as a written word. Humans have been using written languages to store and transmit knowledge for thousands of years. Advances in the production of documents can be seen from the invention of the writing itself, to the re-engineering of the printing press,

INTO JAPANESE

テキスト分析では、書かれた単語として保存されている知識を使用できます。 人間は何千年もの間、知識を保存し伝達するために書かれた言語を使用してきました。 文書の製作の進歩は、筆記自体の発明から、印刷機の再設計まで見て取ることができ、

BACK INTO ENGLISH

In text analysis, you can use the knowledge stored as a written word. Humans have been using written languages to store and transmit knowledge for thousands of years. Advances in the production of documents can be seen from the invention of the writing itself to the redesign of the printing press.

INTO JAPANESE

テキスト分析では、書き言葉として保存されている知識を使用できます。 人間は何千年もの間、知識を保存し伝達するために書かれた言語を使用してきました。 文書の製作の進歩は、筆記自体の発明から印刷機の再設計まで見ることができる。

BACK INTO ENGLISH

In text analysis, you can use knowledge that is stored as a written word. Humans have been using written languages to store and transmit knowledge for thousands of years. Advances in document production can be seen from the invention of writing itself to the redesign of printing presses.

INTO JAPANESE

テキスト分析では、書かれた単語として保存されている知識を使用できます。 人間は何千年もの間、知識を保存し伝達するために書かれた言語を使用してきました。 文書制作の進歩は、筆記自体の発明から印刷機の再設計まで見られる。

BACK INTO ENGLISH

In text analysis, you can use the knowledge stored as a written word. Humans have been using written languages to store and transmit knowledge for thousands of years. Advances in document production range from the invention of writing itself to the redesign of printing presses.

INTO JAPANESE

テキスト分析では、書き言葉として保存されている知識を使用できます。 人間は何千年もの間、知識を保存し伝達するために書かれた言語を使用してきました。 文書制作の進歩は、執筆そのものの発明から印刷機の再設計まで幅広い。

BACK INTO ENGLISH

In text analysis, you can use knowledge that is stored as a written word. Humans have been using written languages to store and transmit knowledge for thousands of years. Advances in document production range from the invention of writing itself to the redesign of printing presses.

INTO JAPANESE

テキスト分析では、書かれた単語として保存されている知識を使用できます。 人間は何千年もの間、知識を保存し伝達するために書かれた言語を使用してきました。 文書制作の進歩は、執筆そのものの発明から印刷機の再設計まで幅広い。

BACK INTO ENGLISH

In text analysis, you can use the knowledge stored as a written word. Humans have been using written languages to store and transmit knowledge for thousands of years. Advances in document production range from the invention of writing itself to the redesign of printing presses.

INTO JAPANESE

テキスト分析では、書き言葉として保存されている知識を使用できます。 人間は何千年もの間、知識を保存し伝達するために書かれた言語を使用してきました。 文書制作の進歩は、執筆そのものの発明から印刷機の再設計まで幅広い。

BACK INTO ENGLISH

In text analysis, you can use knowledge that is stored as a written word. Humans have been using written languages to store and transmit knowledge for thousands of years. Advances in document production range from the invention of writing itself to the redesign of printing presses.

INTO JAPANESE

テキスト分析では、書かれた単語として保存されている知識を使用できます。 人間は何千年もの間、知識を保存し伝達するために書かれた言語を使用してきました。 文書制作の進歩は、執筆そのものの発明から印刷機の再設計まで幅広い。