与其看到吃不到，还是索性看不到的好。
Data Poisoning Attack: This involves inputting antagonistic training data into the classifier. The most common type of attack we observe is model skew. Attackers pollute training data in this way, making classifiers tilt to their preferences when classifying good data and bad data. The second attack we have observed in practice is feedback weaponization, which attempts to abuse the feedback mechanism to manipulate the system to misclassify good content as abuse (e.g. Competitor's content or part of retaliatory attacks).
十九世纪中叶至二十世纪初的自贡，无疑是中国的庞贝城，亦是封建农业社会包围之中的一座手工业城市，其实中国的资本主义已有此地土生土长、生根发芽，发展得相当成熟。
[Truth] On July 19, the Chinese website of China Daily sent a letter of apology to Fan Bingbing, saying: "This website reprinted the above untrue articles without verification, which violated your reputation. Therefore, the website deleted the untrue article as soon as it received the civil complaint on May 24, 2016, and contacted Baidu, 360 and other search engines to delete the snapshot. Because of the timely handling, the false article did not spread widely. "And apologized to Fan Bingbing:" The relevant responsible persons have been dealt with seriously within the network and similar incidents will not happen in the future. Once again, I would like to express my most sincere apologies to Ms. Fan Bingbing. "
< div id= "showcount" >
