Zoom Says AI Translation Models Can Learn From Stronger Models on the Fly
Zoom introduces a reinforcement learning framework that replaces static preference triplets with continuous, high-quality refinements...
Zoom introduces a reinforcement learning framework that replaces static preference triplets with continuous, high-quality refinements...
ByteDance introduces a product-ready AI live speech translation system that delivers near-human accuracy, real-time voice...
A new ByteDance study provides a roadmap for training small, open models that deliver high-quality...
ETH Zurich and EPFL will release a multilingual, fully open LLM trained on the Alps...
Mistral releases Voxtral, a new family of open-source models for AI speech translation and transcription,...
SAP Translation Hub now includes AI translation, expanding its translation offering to over 2,800 language...
Open-source AI Language Proficiency Monitor backed by the German Government ranks commercial and open LLMs...
Google’s Gemini 2.5 delivers stronger performance in CJK and Indic languages, better output language control,...
Microsoft introduces a voice conversion feature in Azure AI Speech, allowing users to transform recorded...
Zoom’s new Realtime Media Streams feature lets developers access live meeting audio, video, and transcripts,...
Large reasoning models perform well in open-domain AI translation, but struggle with domain-specific terminology, instruction...