Event ·Multi-source

谷歌DeepMind推出手语转文本模型，Pixel 11首发支持

谷歌DeepMind发布大规模多语言手语转文本模型SL2T，Pixel 11系列手机上的Gboard和Live Transcribe应用率先支持美国手语输入。该模型在超过50种手语的10万小时数据上训练，能识别身体多部位同步运动。测试显示，美国手语输入比英语打字更快更自然。模型将手语转为姿势特征点，在设备端完成处理，无需上传原始视频，保护用户隐私。

Current read

谷歌DeepMind把手语转文字做进了Pixel 11，美国手语直接变成文本，比打字还快，而且数据在本地处理，隐私有保障。

5 sources78° AI heatLast updated 2026/8/13 02:22:11

Evidence

5 sources

Primary sourceIT之家

2026/8/13 02:22:11View source

Related source 1The Rundown AI

2026/8/13 15:43:36View source

Related source 2Demis Hassabis

2026/8/13 12:53:42View source

Related source 3techcrunch

2026/8/12 14:20:33View source

Related source 4爱范儿

2026/8/13 00:04:10View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story