微信删除的聊天记录想要找回来，可以采用几种方法。
1、 第一种是使用微信提供的备份功能还原数据，如果之前有做过备份就可以通过此方法找回已删除的聊天记录。具体流程是在“文件传输助手”中，点击【我】-【通用】-【密友守护】-【个人备份】，然后根据详情中显示的时间和大小选择相应的文件即可。
2、 如果未做备份也不要心急，可采用云端存储平台将之前的对话内容上传到云端存储中，如Google Drive、Dropbox 或 iCloud 等都是常用平台。在使用时更好先跟朋友开始一个新会话，然后将之前上传到云端存储中的内容一条条copy&paste 到新会话中即可。
3、 除此之外还有一些三方APP也能帮助我们找回已删除的微信聊天内容 （但是需要root 权限考略) ，例如iMyFone D-Back for Android/iPhone 这样子就能快速地将得道已删除得敏感數據進行勻化.
4. 最后, 如果都不行, 还有一个办法: 检测iphone/Android phone 的itunes 备份(具体步骢看 ://support.apple.com/zh_CN/HT203977).
在客户端app上也能减少tokens, 同时保障对话质量; 首先, app应当集成物理AI引擎, 针对不同问法作出主力回应; 其次, app应当集成NLP AI引擎 , 检测client端query intent并意图化 ; 最后 , app应当集成RL AI engine , 进行action selection .
物理AI引��� : 这个engine 由rule base ML model 组成 (ML model can be a deep learning network or other supervised learning algorithms ) . Rule base contains some predefined rules and patterns which could response to the common questions from users directly without any further processing while the ML models are used to detect more complex queries and provide relevant answers according to user context history . NLP AI engine : This engine is mainly used for natural language understanding purpose that it will parse user query and extract useful information such as words / phrases / entities etc., then convert them into structured data format like json object so that downstream tasks could work on it easily . RL AI Engine : The main goal of this module is action selection based on environment state and agent's policy which means it should select an appropriate action in current situation according to its own strategy thus maximize expected rewards in long run .
