" Labeling sound is harder than labeling images because sound is invisible and ephemeral."

“标记声音比标注图像更难，因为声音是无形而且短暂的。”

We downloaded nearly a billion images and used crowdsourcing technology like the Amazon Mechanical Turk platform to help us to label these images.

我们下载了数以百万计的影像，并且使用如Amazon Mechanical Turk 这样的群众外包平台，来协助我们处理及分类这些照片。

You see, AI is evaluated using benchmarks, standardized tests that focus on a specific task, like labeling an image or predicting the rest of a sentence.

您会看到， 人工智能是使用基准进行评估的，标准化测试专注于特定任务，例如标记图像或预测句子的其余部分。