Design of Artificial Intelligence Converged Media Experimental System

Converged media is increasingly becoming a hot technology which integrates multiple media forms due to the development of the 5G mobile communication technology. With up to 10 Gb/s theoretical downstream speed, converged media, including text, ultra-high definition video, 3D audio and mobile applications, are gradually becoming an important development direction of media technology in China. In this paper, a deep learning based artificial intelligence converged media experimental system is designed and built. Specially, we study the methods and technologies of text generation, processing and summarization in converged media system, with emphasis on the webpage content crawling, speech to text technology application and automatic text summarization. To deal with redundant contents, we present Counting Bloom Filter combined Shingling algorithm (CBFCS algorithm) to filter URLs and near-duplicates text. Scheduling Algorithm is used among multi- servers to deal with a large amount of AI tasks. Extensive experiments confirm the effectiveness of our methods.