The Application of Internet of Things and Online Video Recognition in the Cultivation of Intelligent Secretarial Talents

In 2012, the Ministry of education listed secretarial science as a special undergraduate major, which is a qualitative leap in the development of higher Secretarial Education. The traditional training of secretarial talents is mainly oriented to enterprises, government organs and news media, so the training program is more inclined to Chinese major. In this paper, a frame sampling framework based on Multi-Agent Reinforcement Learning is proposed, which is used to sample key frames in long video and improve the accuracy of video recognition. The method proposed in this paper can reduce the cost of computation and reduce the number of queries of non-target attacks by more than 28%.