Network Security Internet Technology Development Database Servers Mobile Phone Android Software Apple Software Computer Software News IT Information

In addition to Weibo, there is also WeChat

Please pay attention

WeChat public account

Shulou

The launch of MathGPT, the first 100 billion-level large model specially built for mathematics in China, opens the public test.

2025-03-28 Update From: SLTechnology News&Howtos shulou NAV: SLTechnology News&Howtos > IT Information >

Share

Shulou(Shulou.com)11/24 Report--

On August 24, in the live broadcast of the 20th anniversary of Hao's future, CTO Tian Mi announced the official launch and public test of the future 100 billion-level model MathGPT in the field of mathematics. From now on, users can apply for a free trial with a registered account through the official website (www.mathgpt.com).

In May this year, good Future announced that it was developing a self-developed mathematical model, named MathGPT. MathGPT is a large model in the vertical field of mathematics, which is aimed at mathematics enthusiasts and scientific research institutions all over the world, with problem-solving and problem-solving algorithms as the core. It is also the first large model specially built for mathematics in China.

When users use MathGPT, they can upload math problems by words or pictures, and they can get conversational solution feedback, and they can also randomly generate math problems and give solutions by the system through the button of "one question at random". Currently, MathGPT supports PC and mobile experiences in both Chinese and English versions.

Leading mathematical problem-solving ability

According to Tian Mi, MathGPT brings together the accumulation of education, teaching and research data for many years in the future, focusing on the field of mathematics. The framework of training, reasoning and deployment of hundreds of billions of large models gives the model a strong ability. Through high-quality education data, to achieve topic calculation, explanation, question and answer and other multi-task continuous training and supervised fine-tuning, showing excellent performance. In addition, with the help of human feedback alignment, the comprehensive literacy of the model will be further improved. MathGPT has obvious advantages in problem-solving accuracy, stability and user experience.

MathGPT's official website shows that MathGPT's mathematical computing ability has covered math problems in primary, junior high and senior high schools, and the types of questions include calculation, application, algebra and other types. It can also ask questions, but there is no question-and-answer interaction outside mathematics.

▲ MathGPT Technical report MathGPT Technical report shows that among the test results of six open mathematical evaluation collections, including CEval-Math, AGIEval-Math, APE5K, CMMLU-Math, College entrance examination Mathematics and Math401, the Future MathGPT has scored the highest scores in many tests. At the same time, MathGPT also performed well in C-Eval 's junior and senior high school general test sets.

▲ MathGPT in the C-Eval list of junior and senior high school grades in terms of problem-solving stability, friendly explanation, MathGPT based on a large number of famous teachers in the process of problem-solving data for model training, the model solving steps are professional and clear.

Taking a series of questions as an example, the answer given by MathGPT consists of three parts: "analysis", "detailed explanation" and "finishing point", which is more detailed than the rough explanation of the general model. "Analysis" provides the idea and way of thinking of the problem, and helps users to understand the problem better. "detailed explanation" gives the specific calculation method and answer, and finally prompts the examination points, difficulties and key points of the question. Help users review and reflect on the intention of the question, and draw inferences from examples.

For users, the study of mathematical problems lies not only in getting the answer itself, but also in the problem-solving principle and logic behind the answer. Compared with other general models, MathGPT can not only solve problems with higher accuracy, but also analyze and explain the answers more clearly, and better meet the core needs of users to use AI products to solve mathematical problems.

At the same time of the release of MathGPT, a representative and challenging set of mathematical task evaluation has been updated on the official website for artificial intelligence experts and mathematics enthusiasts around the world to experience and evaluate. Tian Mi said that he hopes that MathGPT can play a greater role in the field of mathematics education, and that in the future, he is willing to share the experience and methods of research and development of hundreds of billions of large models based on large-scale and high-quality content with the industry, and make common progress with the industry.

Good future AI will rise abruptly based on its accumulated strength

Driven by the wave of AI, a number of technology companies have announced the launch of universal large language model products this year, but the good future has chosen another direction, not based on fine-tuning the existing large language model and interface calls, not doing universal big language model, but in-depth research and development of large models in the vertical field of mathematics, committed to building independent, stable, sustainable, high-quality mathematical solutions.

The general model pays more attention to literature than science, but it has obvious shortcomings in mathematical problem solving, explanation, Q & An and recommendation. On the other hand, mathematical reasoning ability is very important on the way to general artificial intelligence, and many large companies around the world are doing research in this area.

"A good future has 20 years of accumulation in mathematical data and business, a large amount of educational data accumulation and the ability to continuously produce educational data, so choose to do this difficult and correct thing." Tian Mi said that the good future hopes to use his years of accumulation in mathematics and AI to do a good job in the mathematical foundation of the AI model era.

As early as 2017, the AI lab artificial Intelligence Lab was established in the future. With the help of intelligent education, artificial intelligence, open and innovative platform, AI lab won 16 championships and 6 runners-up in all kinds of top academic conference competitions, and published nearly 100 high-level academic papers in international journals and conferences.

In 2019, the Ministry of Science and Technology announced that it would become the first batch of members of the "national team" of artificial intelligence in the education industry, relying on the construction of a new generation of open and innovative platform for artificial intelligence in the future, with years of in-depth research in the field of artificial intelligence. Over the years, the good future, driven by the major needs of the education industry, has built a national educational science and technology innovation platform for education-oriented artificial intelligence algorithm capabilities, application solutions, basic software and hardware systems and open source open services.

The good future is also actively participating in promoting the construction of the large model standard system. As a core unit, it has participated in a series of large model national standards organized by the national artificial intelligence standardization group, and a series of group standards of "large model pre-training model technology and application evaluation methods" led by China Institute of Information and Communication. And the compilation of a series of standards of "General Model of Education" led by the Educational Information Technology Standards Committee of the Ministry of Education and the National Information Technology Standardization Technology Committee.

Recently, the good Future is being used as the lead unit to work with leading scientific research institutions, universities and enterprises such as China Institute of Information and Communication, Fudan University, iFLYTEK, Baidu and other industry leading scientific research institutions, universities and enterprises to comprehensively evaluate the ability of large educational models from the dimensions of scene coverage, application effectiveness and reliable service, so as to provide reference and guidance for the application of educational models.

Using AI to realize large-scale teaching in accordance with their aptitude

With the rise of large language model, how to use AI technology to serve various industries is the focus of social attention. The education industry is one of the earliest industries to lay out the field of AI, and much attention has been paid to what changes AI can bring to the educational ecology.

"AI brings the opportunity to redefine the education industry, and large model technology makes it possible to teach students in accordance with their aptitude on a large scale." Tian Mi said that over the past 20 years, the good future has been exploring personalized learning, from offline small classes to online large classes, and then to AI classes, the form is constantly evolving, but the teaching content is always fixed, there is less interaction between students and teachers, and the granularity can only reach the topic level.

Tian Mi believes that the essence of the large model is a more efficient way to learn knowledge from data and apply it. With the blessing of AI ability, the new learning way of "students' self-study + AI question answering" has become widely possible. The threshold and cost for learners to obtain high-quality teaching content is reduced, and the personalized and refined degree of teaching content continues to increase, which can achieve thousands of people's AI teaching and Q & A tutoring, and each student can get the most suitable learning content.

Based on MathGPT, the good future will continue to explore the way of learning under the AI environment, better serve the global learners and mathematics enthusiasts, share the experience with the industry in time, and promote the positive change of educational science and technology through AI technology.

According to Tian Mi, with the smooth progress of the public test, the problem-solving ability of MathGPT will be continuously improved, and the product-level application based on MathGPT is also being accelerated and will be released in the near future.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.

Views: 0

*The comments in the above article only represent the author's personal views and do not represent the views and positions of this website. If you have more insights, please feel free to contribute and share.

Share To

IT Information

Wechat

© 2024 shulou.com SLNews company. All rights reserved.

12
Report