Revolutionary innovation from Google Gemini! 'When given a grammar book...'

Google has introduced its new artificial intelligence model, Gemini 1.5, and announced that this model can process a 1-hour video at once. It was stated that Gemini 1.5 has been released for developers and enterprise users.

Google announced that improvements have been made in many areas in Gemini 1.5. While Gemini 1.5 Pro performs at the same level as the company's previously released high-end model, Gemini Ultra, it has achieved an 87% improvement over Gemini 1.0 Pro in benchmark tests.

It was stated that this progress was achieved thanks to a technique known as "Mixture of Experts" or MoE. In this technique, instead of processing everything every time a query is sent, only a portion of the model is activated. This approach makes the model faster for users while also making it more efficient for Google.

"IT WAS ABLE TO TRANSLATE"

Turkish theoretical computer scientist and computer science professor Cem Say evaluated this revolutionary innovation of Gemini 1.5.

Say noted the following:

"According to what Google says, the Gemini 1.5 large language model they announced today was able to perform translations on par with humans who use that source when given a grammar book of a little-known language."

Google'ın dediğine bakılırsa bugün duyurdukları Gemini 1.5 büyük dil modeli, az bilinen bir dilin gramer kitabı verildiğinde o kaynaktan yararlanan insanlar ayarında çeviri yapabilmiş. https://t.co/mdS4PtKJJV

— Cem Say (@say_cem) February 15, 2024

News Source: 12punto