Necessary Always Active
Necessary cookies are required to enable the basic features of this site, such as providing secure log-in or adjusting your consent preferences. These cookies do not store any personally identifiable data.
|
||||||
|
||||||
|
||||||
|
Published on
February 21, 2025
5 min read
DeepSeek to Open-Source its Codes as Universities in China Create AI Courses Based on its Models
Universities have started introducing DeepSeek courses in China, Reuters has reported. The Chinese AI startup’s R1 model attracted global attention last month and has been termed as a Sputnik moment for China.
Range of Universities
AI curriculum development in China comes at a time when the Chinese government is looking to boost tech and scientific innovation in universities and schools. This move is aimed at developing new growth sources for the second largest economy in the world.
This week, Shenzhen University said it will introduce an AI course based on DeepSeek. The university, which is based in Guangdong province in southern China, says the course will expose students to key technologies while enabling them to learn about privacy, security, ethics among others. The course is designed to explore ways of striking a balance between tech innovations and ethical norms.
Shanghai-based Jiao Tong University has also introduced
models to upgrade its AI learning tools in the course it offers. Chinese University, Zhejiang said it has already introduced specialized DeepSeek courses earlier this month. Zhejiang is located in eastern China. Another Chinese institution that is leveraging DeepSeek is Renmin University of China. The university says it is applying DeepSeek in multiple fields, powering its research and teaching.
China’s AI Strategy
Introduction of DeepSeek courses is part of China’s broader long-term strategy to lead in emerging technologies. In January of this year, the Asian country released the initial national action plan that’s intended to build a strong education nation by the year 2025.
Reports show that in its plan for AI education in China, the Asian giant is looking to establish a high-quality system of education that’s accessible and ranks among the best in the globe.
DeepSeek has been praised by Silicon Valley executives and engineers in tech companies for its V3 and R1 models. The engineers say the models match some of the most advanced models developed by Meta and OpenAI.
Earlier this week, DeepSeek founder, Liang Wenfeng was among the executives from the Chinese tech industry who participated in a unique meeting with President Xi Jinping. DeepSeek triggered a global selloff last month after it launched its R1 reasoning model. The new model rivaled AI systems developed by top US AI companies in performance and development cost.
Open-Source Code
As universities across China embrace DeepSeek, the AI startup has said that it will be making codes of its models publicly available. This recent move is part of the startup’s commitment to open-source AI. In a post shared on X, the Chinese company said it will open-source five code repositories this coming week. DeepSeek termed the move as small, but one that reflects sincere progress.
“These humble building blocks in our online service have been documented, deployed and battle-tested in production.” DeepSeek’s post read.
DeepSeek’s commitment to open-source its AI models has set it apart from other Chinese AI firms that lean towards closed-source models like their US rivals. In a rare interview conducted last year, Liang Wenfeng said that his company was not prioritizing commercializing their AI models. At the time, the DeepSeek founder alluded to the soft power that underlies open-source applications.
“Having others follow your innovation gives a great sense of accomplishment. In fact, open source is more of a cultural behavior than a commercial one, and contributing to it earns us respect,” Wenfeng said.
DeepSeek’s open-source code will make available the infrastructure that the company uses to support AI models that the company has shared publicly. The company released a new algorithm known as Native Sparse Attention on February 18.
This latest algorithm is designed to make inference and long-context training more efficient. Last month, DeepSeek expanded its user base significantly. As of January 11, it was the most popular AI assistant in China with over 22.2 million active daily users.
