Paradigm of artificial intelligence in big data era
1. School of Computer Science and Technology, Nantong University, Nantong, Jiangsu 226019, China; 2. Nantong Research Institute for Advanced Communication Technologies, Nantong, Jiangsu 226019, China; 3. Cheento Information Technology Co., Ltd., Beijing 100080, China
Abstract：To solve the problems of high cost for knowledge acquisition with poor quality and only simulating low-level intelligence in traditional artificial intelligence paradigm, inspired by the full sample statistic in big data, the approximate searching and the related reasoning, 4 AI paradigms were proposed with perception, data, brain science and cognition as center, respectively. The AI paradigm with perception as center was used to study the distributed and loosely coupled main body to use knowledge, skill and information in coordination and achieve the respective or global goal or plan. The data-centric AI paradigm was formed from quantitative change to qualitative change data, and the core was the data. The AI paradigm with brain science as center was used to make comprehensive research platform based on information and communication technologies. The research purpose of AI paradigm with cognition as center was to use big data to solve the problem of "mind". The paradigm determines the way to realize the AI, and different paradigms have their own characteristics and complement each other. Any one paradigm can't be unified in one format, and the AI paradigm is attributed to 4 paradigms of scientific research.
2017, Vol. 38 
