Achieving Human Parity on Automatic Chinese to English News Translation
read the original abstract
Machine translation has made rapid advances in recent years. Millions of people are using it today in online translation systems and mobile applications in order to communicate across language barriers. The question naturally arises whether such systems can approach or achieve parity with human translations. In this paper, we first address the problem of how to define and accurately measure human parity in translation. We then describe Microsoft's machine translation system and measure the quality of its translations on the widely used WMT 2017 news translation task from Chinese to English. We find that our latest neural machine translation system has reached a new state-of-the-art, and that the translation quality is at human parity when compared to professional human translations. We also find that it significantly exceeds the quality of crowd-sourced non-professional translations.
This paper has not been read by Pith yet.
Forward citations
Cited by 9 Pith papers
-
CodeBLEU: a Method for Automatic Evaluation of Code Synthesis
CodeBLEU improves correlation with human programmer scores on code synthesis tasks by adding syntactic AST matching and semantic data-flow matching to the standard BLEU n-gram approach.
-
PheMT: A Phenomenon-wise Dataset for Machine Translation Robustness on User-Generated Contents
PheMT is a phenomenon-wise dataset created to evaluate NMT robustness against linguistic phenomena in Japanese-English UGC translation, with experiments showing major performance drops on certain phenomena.
-
Forward-Backward Decoding for Regularizing End-to-End TTS
Forward-backward decoding with divergence regularization and bidirectional decoder improves end-to-end TTS robustness and naturalness by addressing exposure bias via joint L2R/R2L training.
-
Findings of the First Shared Task on Machine Translation Robustness
The first shared task on MT robustness received 23 submissions showing up to +22.33 BLEU gains on noisy Reddit data, with strong human-BLEU correlation.
-
Translationese in Machine Translation Evaluation
Translationese in MT test sets biases evaluations, supporting exclusion of reverse-created data, re-evaluation of human-parity claims, and power analysis for reliable significance testing.
-
Metaphors in Literary Post-Editing: Opening Pandora's Box?
Post-editors changed one in three metaphors in NMT and LLM outputs for literary texts, rated quality poor, and found post-editing more laborious than original translation.
-
An Explainable Approach to Document-level Translation Evaluation with Topic Modeling
A topic-modeling framework measures document-level thematic consistency in translations by aligning key tokens across languages with a bilingual dictionary and scoring via cosine similarity, providing explainable insi...
-
Massively Multilingual Neural Machine Translation in the Wild: Findings and Challenges
A single multilingual NMT model for 103 languages trained on 25B examples demonstrates transfer learning benefits for low-resource languages.
-
Survey on reinforcement learning for language processing
This survey reviews reinforcement learning applications to natural language processing problems, especially conversational systems, including problem descriptions, suitability of RL, advantages, limitations, and promi...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.