Home » Expert: Attention to large-scale language model technology will accelerate the realization of general artificial intelligence-News-Science Network

Expert: Attention to large-scale language model technology will accelerate the realization of general artificial intelligence-News-Science Network

by admin


Expert: Attention to large language model technology will accelerate the realization of general artificial intelligence


Chinanews.com, Guangzhou, March 26th (Reporter Cai Minjie) can create poems, draw pictures, and perform medical diagnosis…Recently, conversational large-scale language model technologies represented by ChatGPT and MOSS have become popular all over the Internet, triggering extensive research in various fields. focus on. Artificial intelligence has brought the Internet into a new stage.

“The large-scale language model platform provides new momentum for the development of social governance, smart finance, medical health and smart education.” Qiu Xipeng, a professor at Fudan University and a professor at the AI ​​Model Algorithm Research Center of Pazhou Laboratory (Huangpu), held the first Guangdong-Hong Kong-Macao conference in Guangzhou on the 25th. The award ceremony of the Greater Bay Area (Whampoa) International Algorithm Computing Competition said that the emergence of large-scale language models will accelerate the realization of general artificial intelligence.

Statistics show that the scale of China’s digital economy has ranked second in the world for many years in a row, and new application scenarios and growth drivers continue to emerge. Algorithmic calculation examples are strategic resources and new production factors in the era of digital economy. Hurun Report Chairman and Chief Research Officer Hurun said at the award ceremony that the artificial intelligence industry most closely related to algorithms, China’s core industry, has a scale of more than 400 billion yuan, an increase of more than six times in three years. At the same time, the scale of China’s computing power industry has grown rapidly, with an average growth rate of more than 30% in the past five years.

Not long ago, Qiu Xipeng’s team released MOSS, the first large-scale language model in China, which has attracted much attention. Qiu Xipeng said that compared with ChatGPT, MOSS parameters are relatively small at this stage, and the research and development costs are relatively low. It can be used in vertical industries with a small amount of knowledge and only about 10 billion parameters. However, due to its small capacity and limited knowledge capacity, it needs to be improved in answering questions and logical reasoning ability. Later, it can expand the knowledge in this field and enhance the reasoning ability by expanding the parameter scale or accessing the knowledge base in the vertical field.

What are China’s own advantages at the moment? And how to catch up? Qiu Xipeng believes that China has the ability to develop its own large-scale language model. “The current level of ChatGPT is not too difficult for us to catch up technically, but because it is also developing dynamically, we need to make more efforts to catch up.” Qiu Xipeng said, “We can use domestic high-level talents and massive applications. The advantage of the scene, improve the ability of the underlying algorithm and achieve catch-up with application-driven innovation, and promote the development of large-scale Chinese language models.”

In Qiu Xipeng’s view, the establishment of large-scale language model technology requires multi-party exploration and research. After exploring its technical route, a large number of engineers are needed to realize it. “We need to improve the innovation ecology or mechanism, and cultivate an atmosphere that encourages scientific research and innovation.” He said.

“China’s artificial intelligence market is relatively active and has rich scenarios. It is necessary to establish a large-scale Chinese language model.” Zhu Jun, a professor at Tsinghua University and a professor at the AI ​​Model Algorithm Research Center of Pazhou Laboratory (Whampoa), said that in the field of knowledge annotation, China has high-end knowledge Talent bonus, in the process of dynamic catching up, it is necessary to form a high-quality team of engineering optimization and scientific research exploration to carry out collaborative research, which is more beneficial to resource utilization and research speed.

Zhu Jun said that mainland China has established many artificial intelligence computing centers. “In the long run, we need to master the underlying algorithm capabilities. But for now, we can do new model algorithm innovation first.” He said.

It should be pointed out that if the road of artificial intelligence is to go far, it is necessary to establish red line constraints. “Because the artificial intelligence model iterates very quickly, ethical issues are very important.” Qiu Xipeng suggested that laws and regulations can be used to restrict artificial intelligence products and eliminate some algorithms through the model side to remove harmful information.

Overseas leading research institutions have recently begun to significantly shift to general artificial intelligence based on large models. “The ultimate goal of the development of artificial intelligence is to achieve general artificial intelligence, which is the same as the image of AI in science fiction movies. It has the same ability as humans, especially the general learning ability to complete various tasks.” Qiu Xipeng said, “I am confident about this. I am optimistic, because according to the capabilities currently demonstrated, we are not far away from the era of general artificial intelligence. The next step is to consider connecting the large model to the real world. Through embodied learning, cross-modal learning, etc., let the large model and The real world ‘aligns’.” (End)

Special statement: The reprinting of this article is only for the purpose of disseminating information, and does not mean representing the views of this website or confirming the authenticity of its content; if other media, websites or individuals reprint and use from this website, they must retain the information indicated on this website “Source”, and take legal responsibilities such as copyright; if the author does not want to be reprinted or contact us for reprinting fees, please contact us.

See also  "Working emperor" Tang Jun returns to the public's vision and joins the two major shareholders of Taoli Bread to join Divison

You may also like

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.

This website uses cookies to improve your experience. We'll assume you're ok with this, but you can opt-out if you wish. Accept Read More

Privacy & Cookies Policy