Experts Weigh In on DeepSeek AI Translation Quality
DeepSeek delivers strong multilingual translation at low costs, but language industry experts remain cautious, calling...
Maria Stasimioti
Array
(
[post_type] => post
[posts_per_page] => 12
[paged] => 17
[post_status] => publish
[author] => 2795
)
New Dataset for Low-Resource Kenyan Language Tackles AI Exclusion
Researchers use community engagement and crowdsourcing to build corpora for low-resource Kenyan languages to support...
Apple Teams Up with Italian Researchers for ‘Entity-Aware’ AI Translation
The Sapienza NLP group and Apple collaborate on a shared task to tackle named entity...
How Microsoft Wants to Address Gender Bias in AI Speech Translation
Microsoft researchers show how to mitigate speaker gender bias in large-scale speech translation systems, leveraging...
How Amazon Aims to Improve Lip-Sync in AI Dubbing
Amazon's new audio-visual speech-to-speech translation framework improves AI dubbing by ensuring precise lip-synchrony without altering...
Sony Aims to Improve AI Translation for Indian Language Entertainment Content
Sony Research India’s CASAT framework blends context awareness and style adaptation to elevate entertainment translations...
‘EmoDubber’ to Deliver Emotional Control in AI Dubbing, Researchers Hope
Researchers unveil EmoDubber, an AI dubbing system offering precision in lip-sync, and dynamic control over...
Research Pits Traditional Machine Translation Against LLM-Powered AI Translation
Researchers urge LSPs to evaluate workflows carefully before replacing traditional domain-specific machine translation models with...
Are Large Language Models Ready for Legal Translation?
A new study highlights the uncertainty surrounding the performance of large language models in translating...
Meet QueEn, a Large Language Model for Quechua-English Translation
The University of Georgia introduces QueEn, a large language model designed to bridge communication gaps...
GPT-4 vs. Human Translators Cont’d
Researchers benchmark GPT-4's translation capabilities against human translators and claim the LLM matches junior-level performance...