METHOD AND SERVER FOR TRAINING MACHINE LEARNING ALGORITHM FOR TRANSLATION Russian patent published in 2023 - IPC G06F40/58 

Abstract RU 2790026 C2

FIELD: machine learning.

SUBSTANCE: group of inventions relates to the field of machine learning; it can be used for translation between a rare language and a target language. The method contains stages of: obtainment by a server of a certain real example having the first sentence on one language of a group of languages and the corresponding second sentence on the target language, which is a real translation of the first sentence; generation by the server applying a transliteration function of the first synthetic sentence, thereby generating a synthetic real example having the first synthetic sentence and the corresponding second sentence; obtainment by the server of the third sentence on the target language; generation by the server of the fourth sentence on the rear language for the third sentence, using back translation, wherein the fourth sentence and the third sentence form a certain artificial example; generation by the server applying the transliteration function of the fourth synthetic sentence based on the fourth sentence for the mentioned certain artificial example, thereby generating a synthetic artificial example having the fourth synthetic sentence and the third sentence; training by the server of a translation model based on the synthetic real example and the synthetic artificial example.

EFFECT: increase in the quality of translation from a rear language to a target language.

22 cl, 8 dwg

Similar patents RU2790026C2

Title Year Author Number
METHOD AND SERVER FOR TRAINING MACHINE LEARNING ALGORITHM FOR TRANSLATION 2020
  • Dvorkovich Anton Aleksandrovich
  • Komarov Ivan Sergeevich
RU2789796C2
METHOD AND SERVER FOR PERFORMING PROBLEM-ORIENTED TRANSLATION 2021
  • Emelyanenko Dmitry Viktorovich
  • Ryabinin Maksim Konstantinovich
RU2820953C2
METHOD AND SERVER FOR PERFORMING CONTEXT-SENSITIVE TRANSLATION 2021
  • Golovanov Ilya Aleksandrovich
  • Ivanov Georgy Viktorovich
  • Noskov Aleksey Anatolevich
RU2812301C2
METHOD AND SERVER FOR RANKING DIGITAL DOCUMENTS IN RESPONSE TO REQUEST 2020
  • Dormidontov Sergey Igorevich
  • Tyamgin Ivan Anatolevich
RU2818279C2
METHOD AND SERVER FOR TRAINING MACHINE LEARNING ALGORITHM IN TRANSLATION 2020
  • Dvorkovich Anton Aleksandrovich
  • Kovarsky Boris Andreevich
RU2770569C2
DISTRIBUTED LEARNING MACHINE LEARNING MODELS FOR PERSONALIZATION 2018
  • Kudinov Mikhail Sergeevich
  • Piontkovskaya Irina Igorevna
  • Nevidomskii Aleksei Yurievich
  • Popov Vadim Sergeevich
  • Vytovtov Petr Konstantinovich
  • Polubotko Dmitry Valerievich
  • Malyugina Olga Valerievna
RU2702980C1
METHOD AND SYSTEM FOR TRANSLATING SOURCE SENTENCE IN FIRST LANGUAGE BY TARGET SENTENCE IN SECOND LANGUAGE 2017
  • Gubanov Sergey Dmitrievich
  • Dvorkovich Anton Aleksandrovich
  • Kovarsky Boris Andreevich
  • Nokel Mikhail Alekseevich
  • Noskov Aleksey Anatolievich
  • Frolov Anton Viktorovich
RU2692049C1
SYMBOLS RECOGNITION WITH THE USE OF ARTIFICIAL INTELLIGENCE 2017
  • Chulinin Yurij Georgievich
RU2661750C1
IDENTIFICATION OF IDLE NETWORK CONNECTIONS BASED ON MACHINE LEARNING 2016
  • Li, Xin
  • Yang, Fang
RU2715802C1
METHOD AND SERVER FOR DETERMINING TRAINING SET FOR MACHINE LEARNING ALGORITHM (MLA) TRAINING 2020
  • Dorogush Anna Veronika Yurevna
  • Alipov Vyacheslav Vyacheslavovich
  • Kruchinin Dmitriy Andreevich
  • Oganesyan Dmitry Alekseevich
RU2817726C2

RU 2 790 026 C2

Authors

Dvorkovich Anton Aleksandrovich

Peshkurov Roman Olegovich

Dates

2023-02-14Published

2020-12-22Filed