基于AI对金融类APP防薅羊毛的方法

AI based method to prevent financial app from pulling wool

现有的业务风控系统早已发现有大量的羊毛党利用自动化工具,以非正常手段来完成线上抢促销、套现等恶意行为,从而获取经济利益的行为,但苦于无法实时进行拦截,只能用于事后分析。在促销活动开始时,攻击者会通过自动化工具大量从多点快速登录不同的账号并参与抢促销活动,以极低的代价获取高额利润。而正常用户往往由于服务器过载或者参与速度不及自动化工具,无法享受活动优惠,大幅降低了市场促销活动的效益。为了抵抗自动化工具,只能通过增加更复杂的人工流程来进行判断,比如人脸识别、超复杂验证码;为了让正常用户使用,防止非正常的手段占用的流量与服务器资源,需要投入大量的运维人员时时监控流量,人工封杀、限流等处理,最后还是会处理不及时与大量的误杀,因此急需要解决。

The existing business risk control system has long found that a large number of wool party use automation tools to complete malicious acts such as online sales promotion and cash out by abnormal means,so as to obtain economic benefits.However,due to its inability to intercept in real time,it can only be used for post analysis.At the beginning of the promotion,the attacker will quickly log in different accounts from multiple points and participate in the promotion by means of automated tools,so as to obtain high profits at a very low cost.But the normal users often can’t enjoy the activity discount because of the server overload or the participation speed is not as fast as the automation tool,which greatly reduces the benefit of the marketing promotion.In order to resist the automation tools,we can only judge by adding more complex artificial processes,such as face recognition and super complex verification code;in order to let normal users use and prevent the traffic and server resources occupied by abnormal means,we need to invest a large number of operation and maintenance personnel to monitor the traffic,block and limit the flow manually at all times,and finally we will deal with a large number of problems We need to solve the problem。