Introduction to LanguageX
Champion translation engine + interactive translation , a new generation of translator tools for professional translators
Since neural network machine translation became popular in 2017 , the view that " machine translation replaces human translation" has been seen frequently. As a team with a combination of technical and language backgrounds , LanguageX believes that language services are highly complex intellectual activities that require creativity , and shoulder the important mission of human communication. LanguageX is committed to applying cutting-edge natural language processing technology to the field of professional translation.
In addition to high-quality machine translation engines, LanguageX also allows translators to quickly upload corpus and train their own translation engines , greatly reducing the threshold for translators to use professional machine translation.
In addition, LanguageX provides AI -assisted translation functions such as interactive translation, intelligent error correction, and real-time preview. The translator is in the process of translating. The system recommends translations in real time based on terminology, translation memory, and translator input. Let translators focus on translation , greatly improving translation efficiency and experience.
Win the championship in one fell swoop! LanguageX wins the English-Chinese Championship in the 2022 WMT International Machine Translation Competition
LanguageX function introduction
translation engine
highlights
1. Integrates 13 mainstream MT engines, which can visually compare translation results;
2. Provide mainstream MT engines such as Baidu, Ali, and Huoshan, which are more targeted in medicine, finance, and web translation;
3. Support users to share corpus, participate in the training of MT engines in vertical fields, or train personal MT engines.
insufficient
1. The function of self-built MT engine uploaded by individuals currently only supports Chinese and English language pairs;
2. There are fewer MT engines in the vertical field , and users still need to participate in the joint construction.
Translation Project: Supports file uploads in multiple file formats
translation memory
The memory bank stores the historical translation results. For example, in an article, a sentence is translated from A to B (assuming it is Chinese to English), then the next time you encounter translation A in another article (Chinese In the case of English-to-English translation, there is no need to manually translate again, but the translation results from the memory can be directly used to improve translation efficiency and even affect the calculation of translation costs.
The term base stores a series of terms. For example, Facebook may be translated into any need as Facebook , so the record in the term base is Facebook -> Facebook . Or there is a word that will be translated into different results in different languages, then the records in the term base may be: A -> B (Chinese), A -> C (English). When encountering A during translation , it will automatically complete the translation in combination with the term base, or verify the translation results of A.
AI Toolbox
LanguageX provides two alignment modes - AI alignment version and regular alignment version . The former is suitable for more complex or irregular documents, which are aligned by AI ; the latter is suitable for documents corresponding to more standardized documents, which are segmented according to segments Regular alignment. The " AI Alignment Version " divides the imported text into 27 Chinese-English sentence segments, with a high accuracy rate and no need for manual alignment adjustment. The " rule alignment version " is divided into 9 contrasting sentence segments according to the original text paragraphs. For the convenience of subsequent use, it is necessary to further manually segment the paragraphs for alignment. Compared with before, the imported document " AI alignment version" is more efficient in segmentation, and the number of segmented segments is three times that of the latter.
Select double document/ single document, upload the original translation, select the language direction, and the machine translation engine you want to detect. In order to effectively control the flow of machine translation, you can choose the amount of random inspection and filter the sentences that are too short. Click [ Detect ] to start detection.
After the detection is complete, click the card to enter the details. Click the engine name, switch to view the detection similarity and difference marks in the third column of the table. Support downloading reports to local.
LanguageX uses
URL: LanguageX
New users can try membership for 7 days for free, with 500,000 characters.
