Xiaomi: The sound recognition algorithm ranks first in international performance, and the audio labeling task has made important progress

According to news from this site on September 13, Xiaomi announced today that Xiaomi’s self-developed voice recognition algorithm has made important progress in the audio tagging task.

According to reports, using the audio data of the public data set AudioSet-2M as the audio labeling model of the training set, the score exceeded 50 mAP for the first time. This breakthrough marks that Xiaomi's voice recognition algorithm has achieved international success. Ranked #1 in performance.

Xiaomi said: “In the future, we will continue to explore new heights of technology and use it in a variety of device usage scenarios such as mobile phones, speakers, bracelets, and CyberDog. ", bringing users a more efficient and accurate voice recognition experience."

As previously reported by this site, Xiaomi is also currently working on large AI models, and has combined with Xiaoai classmates to further expand user coverage. Combined with the advancement of voice recognition algorithms, Xiaoai is expected to have a better experience.

Lei Jun revealed in his previous annual speech that in terms of AI, Xiaomi began its layout in July 2016 and was established in April this year. Large model team, with more than 3,000 related teams. The latest 1.3 billion parameter large model of Xiaomi AI has been successfully run locally on mobile phones, and some scenarios can be comparable to the results of running a 6 billion parameter model in the cloud.

Advertising statement: The external jump links (including but not limited to hyperlinks, QR codes, passwords, etc.) contained in the article are used to convey more information and save selection time. The results are for reference only. All articles on the site contain this statement.

The above is the detailed content of Xiaomi: The sound recognition algorithm ranks first in international performance, and the audio labeling task has made important progress. For more information, please follow other related articles on the PHP Chinese website!