Shulou(Shulou.com)11/24 Report--

Thanks to CTOnews.com netizen OC_Formula for the clue delivery! CTOnews.com, January 10, Microsoft recently released an artificial intelligence tool called VALL-E, which can imitate human speech in 3 seconds of audio.

The tool is trained for 60000 hours of English voice data and uses 3-second clips of a specific voice to generate content. Unlike many current artificial intelligence tools, VALL-E can copy the mood and tone of the speaker, even words that the speaker has never said before.

CTOnews.com learned that a Cornell paper synthesized several sounds using VALL-E, and you can listen to the audio synthesized by AI on GitHub.

The researchers point out that in many cases, Vall-E performs better than current text-to-speech models. However, the study also wrote that there are several problems with artificial intelligence models. For example, some words in a text prompt may be inarticulate, completely omitted, or appear twice in the output. In addition, the model is currently difficult to imitate some sounds, especially those with an accent.

Like other new AI technologies, VALL-E also raises concerns about security, ethics and so on. Microsoft has issued an ethical statement about the use of VALL-E, but it is not clear about its future use.

At present, Microsoft Vall-E has not been opened yet. Microsoft has created a Vall-E repository on GitHub, but currently contains only one description file.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.