Google to introduce a 1000-language AI model to beat ChatGPT
Google is expected to introduce numerous AI capabilities for its products in the near future, including the Imagen text-to-image generator, which Gboard for Android is working to integrate.
Google has provided more details about the Universal Speech Model (USM), which the company believes is a “critical first step” towards building an AI language model that can support 1,000 different languages and surpass ChatGPT.
In November of last year, Google announced its intention to develop a language model that could handle 1,000 of the world’s most widely spoken languages and revealed its USM model. The USM is a family of advanced speech models with 2 billion parameters that were trained on 12 million hours of speech and 28 billion text sentences, covering over 300 languages.
According to Google, the USM is currently capable of performing automatic speech recognition (ASR) in over 100 languages, including under-resourced languages such as Amharic, Cebuano, Assamese, and Azerbaijani. The company plans to use the USM as the foundation for a much larger language model.
Additionally, Google is expected to introduce numerous AI capabilities for its products in the near future, including the Imagen text-to-image generator, which Gboard for Android is working to integrate.
“USM, which is for use in YouTube (e.g., for closed captions), can perform automatic speech recognition (ASR) not only on widely spoken languages like English and Mandarin, but also on under-resourced languages like Amharic, Cebuano, Assamese, and Azerbaijani to name a few,” Google said in a blogpost.
Google currently claims that USM supports over 100 languages and will serve as the “foundation” for a much larger system.
