Intellectual hub of the topic
Researchers have created a specialized safety test for language models that accounts for the nuances of Thai language and culture. This project has already been accepted into a major AI workshop.
Boson AI's Higgs Audio v3 recognizes speech in 94 languages, understands emotions, and surpasses competitors in accuracy for key languages.
Japanese company Rakuten has released its new language model, Rakuten AI 3.0, positioning it as the country's largest high-performance AI model and developed with government support.
Mistral has released Small 4, a new compact model that's faster, more accurate, and boasts improved performance across multiple languages, including Russian.
Yandex AI Studio has updated its file search tool, enabling AI agents to work with tables, audio, and video to find information in corporate knowledge bases.
AI: Events
Products
Mixedbread has introduced Wholembed v3 – a multimodal search model that works with text and images in dozens of languages and claims to deliver best-in-class results.
AI: Events
Development
We explore how modern speech recognition systems have learned to extract specific data – phone numbers, addresses, and emails – from conversations on the fly.
AI: Events
Development
We explore how diarization works – the technology that determines who is speaking and when in an audio stream – and why doing it in real time is particularly challenging.
AssemblyAI has released the Universal-3 Pro model, which supports six languages and allows switching between them mid-speech without manual adjustments.
Don’t miss a single experiment!
Subscribe to our Telegram channel —
we regularly post announcements of new books, articles, and interviews.