Google announces AudioPaLM, a large-scale spoken language model
Haebom
1
Created by
Haebom
Created at
Google announces Audio PaLM, a combined result of PaLM-2 and AudioLM, a text-based and speech-based language model
Features an integrated multi-modal architecture capable of simultaneously processing and generating text and speech for applications such as speech recognition and speech-to-speech translation.
AudioPaLM simultaneously utilizes the ability to preserve additional linguistic information such as speaker identity and stress obtained from AudioLM, and the linguistic knowledge of text-based language models such as PaLM-2.
Simply put, the era of simultaneous interpretation and translation is not far off.
The model demonstrates superior performance compared to existing systems in speech translation tasks.
Ability to perform zero-shot speech-to-text translation for multiple languages without input/target language pairs during training
AudioPaLM can implement audio language model functions such as language-to-language speech conversion based on short speech prompts.
📚 Welcome to Haebom's archives. --- I post articles related to IT 💻, economy 💰, and humanities 🎭. If you are curious about my thoughts, perspectives or interests, please subscribe. Would you like to be notified when new articles are posted? 🔔 Yes, that means subscribe. haebom@kakao.com
Subscribe
1
OAKPDNOW
이거 진짜 대박사건이네! AI 가 멀티링구얼 스피킹을 하게 해주면 너무 좋을 듯!! 세계와 경험이 확장되니까! :)
