Yandex launched a service for decrypting archived documents using a neural network


"Yandex” launched the “Search in Archives” service, which contains more than 2.5 million pages of historical handwritten documents with text transcripts. This was reported by the press service of the company.

Handwritten pre-revolutionary documents are deciphered by a neural network. It is noted that experts trained it on a data array of hundreds of thousands of handwritten lines of real texts of the 18th-19th centuries. from the Main Archive of Moscow, as well as tens of millions of generated examples. Before that, they themselves marked up and deciphered historical materials.

With the help of technology, users can find the desired document in the service database with a mention of the last name, locality, etc. Text search is available in the catalog or through the search bar, there are also filters by years, archives, funds and inventories. The document decoded by the neural network is displayed to the right of the original, when you hover over the desired fragment, it will immediately be highlighted on the digital copy of the text, the press service explained.

In addition to the documents of the Main Archive of Moscow, the database also contains texts from the archives of the Orenburg and Novgorod regions. The developers promise that in the future there will be even more documents available in the service.

In November 2022, Yandex will also taught the neural network to find places for complex queries in Yandex.Maps. When processing a request, it analyzes not only the description and keywords, but also other parameters - photos, reviews, and menus.



Source link