Post-editing processing method for correction of wrong words in machine translation

Abstract

The invention relates to the field of machine translation, in particular to a post-editing processing method for correction of wrong words in machine translation. A translated word error correction rule template is introduced into machine learning, a source text, a standard translated text and a machine translated text are compared through the rule template, rules for wrong word correction are extracted from a training set, and the extracted rules are utilized to correct translated word errors in translation of a machine translation system, so that the repeated translated word errors in machine translation are remarkably reduced, and the labor and the time in manual post-editing are reduced. According to the method, through template matching and continuous iteration, the correction rules are extracted to form a rule sequence; the rule sequence extracted by the method is utilized to correct the wrong translated words in the machine translated text continuously, so that the machine translated text continuously approaches the standard translated text, the quality of the translated text of machine translation is remarkably improved, and the method has wide application prospects in all kinds of machine translation systems.

Description

technical field [0001] The invention relates to the field of machine translation, in particular to a post-editing method for correcting wrong words in machine translation. Background technique [0002] Now that the Internet has spread all over the world, people from different nationalities and nationalities can share and exchange information anytime and anywhere; people are eager to obtain all the information on the Internet quickly and smoothly. Therefore, there is a huge market demand for accurate and efficient automatic machine translation between multiple languages ​​in the current and future international atmosphere. However, an Internet multilingual translation system with high performance, powerful functions and high accuracy still needs to overcome many major technical difficulties under the current technical level. With the current level of machine translation, high-quality usable machine translations are still unavailable. The current general way to solve this pr...

AI Technical Summary

Problems solved by technology

Although the research on automatic post-editing based on statistical machine translation has achieved certain results; but there are still many things that are not clear about what happened in statistical machine translation SMT

As far as this post-editing technology is concerned, it is only known that this method can improve the quality of the final translation result, but it is not known which specific post-editing operations are effective (that is, which part of the post-editing operations represents the defects of the machine translation system), which does not Conducive to intuitive analysis of the disadvantages of machine translation

Images