How Question Answering Can Transform AI Translation Evaluation
Two new research papers propose using question answering to evaluate AI translation, challenging how the...
AI translation
From enterprise translation AI to automated localization.
Multi-Agent AI Systems Can Improve Accuracy and Domain Adaptation in AI Translation
ADAPT researcher says multi-agent AI setups with specialized LLMs for translation, review, and final editing...
How Well Can OpenAI’s o3-mini and DeepSeek-R1 Evaluate AI Translation?
A new study examines how well reasoning-enabled large language models evaluate AI translation quality and...
Lessons from AI Translation to Improve Multilingual LLM Evaluation
Recent studies reveal gaps in multilingual LLM evaluation, showing translation alone is insufficient. Researchers call...
Alibaba and Meta Face Off in Simultaneous AI Translation
Alibaba and Meta aim to improve simultaneous AI translation by training large language models for...
The European Commission Launches a Free Web AI Translation Plugin
The European Commission releases WEB-T, a free automatic translation browser plugin meant to help European...
Google Calls for Rethink of Single-Metric AI Translation Evaluation
Google and Imperial College researchers propose evaluating AI translation systems on an accuracy-naturalness plane instead...
AI Translation Improves for Indic Languages. Preserving Sentiment Still a Challenge
A new study finds that while large language models improve translation accuracy for Indic texts,...
How Large Language Models Improve Document-Level AI Translation
University of Zurich and Huawei researchers show how LLMs improve document-level AI translation by preserving...
Does Word-Level Quality Estimation Really Improve AI Translation Post-Editing?
A new study investigates the impact of word-level quality estimation on post-editing. While it may...
Multi-Agent AI Framework Aims to Improve Cultural Adaptation in AI Translation
Researchers from Shahjalal University of Science and Technology and the University of Oklahoma propose a...
How France’s Inria Aims to Improve AI Translation for Low-Resource Languages
Inria improves AI translation for low-resource languages by breaking sentences into simpler phrases, translating them...
Translationese Remains a Challenge for Large Language Models, Study Finds
A new study finds that translationese — unnatural, overly literal translations — remains a persistent...
Unbabel Tackles Metric Bias in AI Translation
Unbabel exposes how using the same metrics for both training and evaluation can create misleading...