微软一直都号称拥有最大的人脸数据库，任何人都可以使用它来训练面部识别算法。
因为媒体的报道引发了关于隐私和道德问题的关注，曾经公开可用的数据集已经悄然消失。
现已删除的数据集中包含超过1000万张面孔，这些面孔来自Flickr等网站，这些网站储存的是根据知识共享许可上传的照片——这意味着许多都可以免费，但可能有版权问题。
其中的许多面孔不属于名人——包括记者和隐私研究人员——并且他们不知道自己的图像被包括在内。
当然，微软并不是唯一一家通过从开放的互联网上抓取照片来组装大型数据集的公司。
调查显示，很多创业公司一直在使用照片收集应用程序暗中收集数百万张面孔，而其他公司则一直在扫描大量的大头照。
Hover a word for its pinyin & meaning; click to pin it.
narration saved locally, offline — from https://download.mandarinbean.com/audio?source=aHR0cHM6Ly90cmFmZmljLmxpYnN5bi5jb20vbGVhcm5jaGluZXNlL01pY3Jvc29mdF9kZWxldGVkX3RoZV9mYWNlX3JlY29nbml0aW9uX2RhdGFiYXNlXy1fNjA2Ni5tcDM&filename=Microsoft-deleted-the-face-recognition-database.mp3&signature=6f7260dd13d2964677fc77e79e22b3b12b24020507501a716c34880a34e72a94
| Chinese | Pinyin | English |
|---|---|---|
| 微软 | Wēi ruǎn | Microsoft Corporation |
| 一直 | yī zhí | straight (in a straight line) / continuously; always; all the way through |
| 都 | dōu | all; both; entirely |
| 号称 | hào chēng | to be known as; to be nicknamed / to be purportedly / to claim (often exaggeratedly or falsely) |
| 拥有 | yōng yǒu | to have; to possess |
| 最 | zuì | (before an adjective or verb) to the highest degree; (the) most ...; -est |
| 大 | dà | big; large; great / older (than another person) / eldest |
| 的 | de | possessive particle (of; 's) / modifying particle (used after an attribute to modify a noun or to form a nominal expression) |
| 人 | rén | person; people |
| 脸 | liǎn | face |
| 数据 | shù jù | data |
| 库 | kù | warehouse / storehouse / (file) library |
| 任何 | rèn hé | any; whatever; whichever |
| 可以 | kě yǐ | can / may / possible / able to |
| 使用 | shǐ yòng | to use / to employ / to apply / to make use of |
| 它 | tā | it |
| 来 | lái | (between two verbs) in order to |
| 训练 | xùn liàn | to train / to drill |
| 面部 | miàn bù | face (body part) |
| 识别 | shí bié | to distinguish; to discern; to identify; to recognize |
| 算法 | suàn fǎ | arithmetic / algorithm / method of calculation |
| 因为 | yīn wèi | because; owing to; on account of |
| 媒体 | méi tǐ | media, esp. news media |
| 报道 | bào dào | report |
| 引发 | yǐn fā | to lead to / to trigger / to initiate / to cause / to evoke (emotions) |
| 了 | le | (completed action marker) / (modal particle indicating change of state, situation now) |
| 关于 | guān yú | pertaining to / concerning / with regard to / about / a matter of |
| 隐私 | yǐn sī | secrets / private business / privacy |
| 和 | hé | and; together with |
| 道德 | dào dé | virtue / morality / ethics |
| 问题 | wèn tí | question / problem / issue / topic |
| 关注 | guān zhù | concern; interest; attention |
| 曾经 | céng jīng | once / already / former / previously / ever / (past tense marker used before verb or clause) |
| 公开 | gōng kāi | open; overt; public |
| 可用 | kě yòng | available for use; usable |
| 集 | jí | to gather / to collect |
| 已经 | yǐ jīng | already |
| 悄然 | qiǎo rán | quietly |
| 消失 | xiāo shī | to disappear; to fade away |
| 现 | xiàn | now |
| 已 | yǐ | already |
| 删除 | shān chú | to delete / to cancel |
| 中 | zhōng | within; among; in |
| 包含 | bāo hán | to contain / to embody / to include |
| 超过 | chāo guò | to surpass / to exceed / to outstrip |
| 万 | wàn | ten thousand / a great number |
| 张 | zhāng | classifier for flat objects, sheet / classifier for votes |
| 面孔 | miàn kǒng | face |
| 这些 | zhè xiē | these |
| 来自 | lái zì | to come from (a place) |
| 等 | děng | and so on; etc.; et al. / (used to end an enumeration) |
| 网站 | wǎng zhàn | website |
| 储存 | chǔ cún | to store / to stockpile |
| 是 | shì | to be (followed by substantives only) |
| 根据 | gēn jù | according to / based on |
| 知识 | zhī shi | knowledge |
| 共享 | gòng xiǎng | to share / to enjoy together |
| 许可 | xǔ kě | to allow / to permit |
| 上传 | shàng chuán | to upload |
| 照片 | zhào piàn | photograph / picture |
| 这 | zhè | this; these |
| 意味 | yì wèi | to mean / to imply / to get a sense of |
| 着 | zhe | aspect particle indicating action in progress or ongoing state |
| 许多 | xǔ duō | many / a lot of / much |
| 免费 | miǎn fèi | free (of charge) |
| 但 | dàn | but; yet; however; still |
| 可能 | kě néng | might (happen) / maybe / perhaps |
| 有 | yǒu | to have; there is |
| 版权 | bǎn quán | copyright |
| 其中 | qí zhōng | among / in / included among these |
| 不 | bù | no; not so |
| 属于 | shǔ yú | to belong to; to fall under (a category) |
| 名人 | míng rén | personage / celebrity |
| 包括 | bāo kuò | to comprise / to include / to involve / to incorporate / to consist of |
| 记者 | jì zhě | reporter; journalist |
| 研究 | yán jiū | research / a study |
| 人员 | rén yuán | staff / crew / personnel |
| 并且 | bìng qiě | and; besides; moreover; furthermore; in addition |
| 他们 | tā men | they; them |
| 知道 | zhī dào | to know; to become aware of |
| 自己 | zì jǐ | oneself; (attributive) one's own |
| 图像 | tú xiàng | image; picture; graphic |
| 被 | bèi | used to indicate passive voice (placed before the doer of the action like "by" in English passive-voice sentences, or, if the doer is not mentioned, before the verb) / (sarcastic or jocular) used to indicate that the following word should be regarded as being in air quotes |
| 在 | zài | (of sb or sth) to be (located) at |
| 内 | nèi | inside / inner / internal / interior |
| 当然 | dāng rán | certainly / of course / without doubt |
| 并 | bìng | (not) at all |
| 不是 | bù shì | is not |
| 唯一 | wéi yī | only; sole |
| 一 | yī | one / single / a (article) |
| 家 | jiā | classifier for families or businesses |
| 通过 | tōng guò | by means of; through; via |
| 从 | cóng | from; through; via / (bound form) retainer; attendant / (bound form) assistant; auxiliary; subordinate / (bound form) related by common paternal grandfather or earlier ancestor |
| 开放 | kāi fàng | to bloom / to open / to be open (to the public) / to open up (to the outside) / to be open-minded |
| 互联网 | Hù lián wǎng | Internet |
| 上 | shàng | (noun suffix) on; above |
| 抓取 | zhuā qǔ | to seize |
| 组装 | zǔ zhuāng | to assemble; to put together |
| 大型 | dà xíng | large / large-scale |
| 公司 | gōng sī | company; firm; corporation |
| 调查 | diào chá | investigation / inquiry / survey / (opinion) poll |
| 显示 | xiǎn shì | to show / to illustrate / to display / to demonstrate |
| 很多 | hěn duō | many; a lot of |
| 创业 | chuàng yè | to begin an undertaking; to start an enterprise |
| 收集 | shōu jí | to gather / to collect |
| 应用程序 | yìng yòng chéng xù | (computing) program / application |
| 暗中 | àn zhōng | in the dark / in secret / on the sly / surreptitiously |
| 数 | shù | several; a few |
| 百万 | bǎi wàn | million |
| 而 | ér | but (not) / yet (not) / (indicates contrast) |
| 其他 | qí tā | other / (sth or sb) else / the rest |
| 则 | zé | (literary) (conjunction used to express contrast with a previous clause) but; then |
| 扫描 | sǎo miáo | to scan |
| 大量 | dà liàng | great amount / large quantity / bulk |
| 头 | tóu | head / the top / end / beginning or end / chief / boss / side / aspect |
| 照 | zhào | photo |