Chinese Robot Challenged CEE Math Papers and Scored 105 Points at Best

Jun 9, 2017 - (Newswire)

- After Google AlphaGo, another artificial intelligence came. A science-and-technology enterprise - Chengdu Zhunxingyunxue Technology Co., Ltd. seated in Chengdu Hi-tech Zone has developed an AI system AI-MATHS. As an examination robot, it challenged math questions of 2017 Chinese College Entrance Examination (CEE) on June 7th.

This examination robot is an AI system developed based on Tsinghua University big data, artificial intelligence, natural language recognition and cutting-edge technologies, with automatic question answering technology as the core. The R&D team comes from a local artificial intelligence research and development company in Chengdu Hi-tech Zone.

After the CEE math exam in the afternoon of June 7th, the examination robot stayed alone in a room with the external network off and invigilated by specialized persons, and after finished reading the questions on e-paper, it began to calculate via the internal server and then gave answer. AI-MATHS finished multiple examination papers in two hours and the answers were then scored by the teachers. AI-MATHS finished the math paper of liberal arts CEE mock examination of Beijing in 22 minutes and scored 105 points, then finished liberal arts CEE national unified paper II in ten minutes and scored 100 points.

AI-MATHS robot programme is led by "Super Brain Plan" of Chinese 863 Program. Lin Hui, director of the large data center at the Suzhou Automotive Research Institute of Tsinghua University, serves as CEO of Zhunxingyunxue. He said, the company's core team involved and led the project Key Technology and System of Elementary Mathematics Problem Solving subsidiary to the National 863 project Key Technology and System of Large Data Based Humanoid Robot Intelligence (i.e. math examination robot).

According to Lin Hui, AI-MATHS applies small-scale training, and in order to achieve quick answer, the team increased the number of questions. AI-MATHS currently has 500 sets of examination, a total of 12,000 questions.

Since then, AI-MATHS will receive large-scale training and flexible expansion ability testing, even involving learning ability in other fields. At present, the technology company has landed AI education products mainly for mathematics of secondary education, and they cover dozens of chapters and 4,000~5,000 knowledge points and testing points, significantly reducing the cost of various models and labor and property costs of manual annotation.

Lin Hui also said that in the future AI-MATHS will not only be limited to education field, but also turn to the domains of finance, medical treatment and transportation, etc.

While the craze for AI has swept the world, China has included it into its national strategy. In March 5th, 2017, AI was first referred in the State Council's work report. As China's AI research is in an outbreak, industrial giants gradually optimize their deployment of AI industrial chain, and start-ups spring up to persistently seek in vertical.

Experts pointed out that the AI industry will undoubtedly become the locus of "unicorns companies", and talent pooling will turn to be an important factor affecting China's AI development.

Lin Hui revealed the company is looking for the global high-end technical personnel for further expansion and the development of the company is strongly supported by the AI-based industries gathered in Chengdu Hi-tech Zone including software development, cloud computing, big data, networking and intelligent manufacture and also by sound talent environment here. Presently, 90% of the company's talented people comes from Chengdu, and particularly, a large number of technical backbones are recruited from University of Electronic Science and Technology of China (UESTC). The chief scientist of UESTC also provided the algorithm and theoretical support for AI-MATHS.

SOURCE Chengdu High-tech Industrial Development Zone

Original Source: https://www.newswire.com/news/chinese-robot-challenged-cee-math-papers-and-scored-105-points-at-best

What Next?

Related Articles

Leave a Reply

Submit Comment