AIExplainer

What is a BLEU?

A metric for evaluating the quality of machine translation

Stands for: Bilingual Evaluation Understudy

BLEU measures how well a machine translation matches a human translation by comparing the overlap in words and word sequences

BLEU is like a grading system for translation quality, where a higher score means the machine translation is closer to the human translation, just like how a higher grade means better performance on a test

Google Translate uses BLEU to evaluate and improve the quality of its translations, ensuring that users receive the most accurate translations possible

BLEU is used to evaluate and compare the performance of different machine translation systems, helping to identify the most accurate and effective ones

BLEU is not a perfect measure, as it can be influenced by factors like language complexity and translation style, and may not always reflect the actual quality of the translation

BLEU was developed in 2002 by researchers at IBM as a way to automatically evaluate machine translation quality

translation quality metric machine translation evaluation metric

Three products for different needs — explore what’s relevant to you.