The college entrance examination robot challenges the college entrance examination mathematics volume for the first time, 2 hours to answer multiple sets of questions

(Original title: The college entrance examination robot will challenge the college entrance examination mathematics volume for the first time and use 2 hours to answer multiple sets of questions.)

News reporter Xu Diwei

Another artificial intelligence came. An artificial intelligence system AI-MATHS developed by Chengdu Hi-tech Zone Technology Co., Ltd. will be used as a "math college entrance examination robot" and will challenge the 2017 college entrance examination mathematics volume on June 7.

This college entrance examination robot is based on Tsinghua University big data, artificial intelligence, natural language recognition and other cutting-edge technology research and development of an artificial intelligence system with automatic problem-solving technology as the core, its R & D team from Chengdu High-tech Zone, a local artificial intelligence research and development company.

On June 2, the relevant person of the R&D team told reporters () that after the completion of the mathematics examination on the afternoon of the 7th, the “entrance robot” will be held alone in a room off the external network. Profession notaries will inspect the input. The electronic test papers of the test papers are used to read the questions and the answers are displayed through the calculation of the internal server. On the day of the challenge, the AI-MATHS will answer multiple high-achieving questions simultaneously within two hours. Afterwards, the teachers will correct the responses.

AI-MATHS is a robot led by the Ministry of Science and Technology of the People's Republic of China (also known as the "brain project") and has been researched for more than four years. Lin Hui, director of the Big Data Center of the Suzhou Research Institute of Tsinghua University, is the CEO of the technology company. He introduced that the company's core team took part in and led the sub-topic "Key Techniques and Systems for Solving Elementary Math Problems" of the National 863 Project "Key Technology and System for Human Intelligence Based on Big Data" (ie, mathematics college entrance examination robot).

The college entrance examination robot AI-MATHS was born in 2014. In May 2016, its R&D team announced that AI-MATHS will participate in the 2017 college entrance examination.

In February 2017, AI-MATHS scored 93 points in the mathematics simulation volume of the college entrance exam alone on the basis of 100 sets of test questions and 12,000 training points. It exceeded the average score of the 2016 mathematics of the Sichuan College Entrance Examination and passed the mid-term evaluation.

In order to train the answer speed, the team has increased the number of questions. Currently, AI-MATHS has accumulated 500 test questions.

However, although artificial intelligence has superior capabilities in computing, reasoning, and storage, the breakthrough lies in having the ability to self-learn and self-optimize. Lin Hui introduced that AI-MATHS solves the problem through a comprehensive logical reasoning platform, rather than learning to store question bank. "It can learn more than 7,000 test sites from elementary school to high school, and the computational power can reach 800 power."

In the past more than one year, the R&D team has conducted in-depth research on this college entrance examination robot in complex logic reasoning, intuitive observation and reasoning, computer algorithms, and deep learning. However, there are still some bugs that need continuous improvement. For example, AI-MATHS lacks Common sense comprehension ability, "you can only guess if you don't understand the topic."

This open challenge to the college entrance examination mathematics volume is the first demonstration of the college entrance examination robot system to the public.

PA Woofer

Woofer Speaker,Pa Subwoofe,Speaker Sub Woofer,96Dbw/M Professional Speaker

NINGBO RFUN AUDIO TECHNOLOGY CO.,LTD , https://www.mosensound.com