Source: Quantum Bit

Image Source: Generated by Wujie AI

Transitioning from AI Pharmaceuticals to the Field of Large Models!

This is what Qi Yuan, former vice president of Ant Group and current director of AI Innovation and Industry Research Institute at Fudan University, is doing.

His company, Infinite Light Years, has now shifted its focus from the AI pharmaceutical field to the development of generative AI and AGI.

Quantum Bit has learned that the company is planning to independently develop the foundation for building large models in the tens of billions to hundreds of billions, ultimately constructing industry applications for the medical, gaming, and education sectors.

At the end of 2021, Qi Yuan, then vice president of Ant Group, returned to academia and became the director of the AI Innovation and Industry Research Institute at Fudan University. In the first half of the following year, Qi Yuan founded Infinite Light Years.

At the time of its establishment, the company positioned itself as an AI pharmaceutical company, primarily engaged in "medical research and experimental development (excluding human stem cells, genetic diagnosis, and treatment technology development and application)."

At that time, Qi Yuan's LinkedIn profile included a description related to "AI for drug design," but this phrase has since been removed from his profile.

The company previously secured a multimillion-dollar investment from Qiming Venture Partners. At Qiming Venture Partners' forum this year, Infinite Light Years was listed as one of its invested projects in generative AI, along with other projects such as Zhipu AI and Xianyuan Technology.

Infinite Light Years' Path to Large Models

In May 2022, Infinite Light Years was established, focusing on AI pharmaceuticals.

Currently, the company has completely closed pharmaceutical-related technical positions on its recruitment platform, and the existing recruitment directions are mostly related to algorithms and infrastructure architecture.

So, after transitioning from AI pharmaceuticals and adjusting its positioning, how will Infinite Light Years focus on the field of large models?

According to market reports, the company currently has a staff size of around fifty to sixty people, with the majority of members coming from Alibaba, as well as members with backgrounds from Microsoft, Tencent, and ByteDance, with a research and development ratio of around 85%.

Regarding the first hurdle for entrepreneurial large models—computing resources, it is rumored that the company has abundant cluster resources and is not worried about this.

The latest news is that the company is planning to independently develop the foundation for building large models in the tens of billions to hundreds of billions, and will also embark on the development of domain-adaptive intelligent agents, enabling machines to solve problems in complex environments by simulating human learning and behavior.

In this regard, the chosen path seems to be somewhat related to the AI Agent trend sparked by projects such as AutoGPT and BabyAGI.

As for the specific differentiated approach that will be adopted, Qi Yuan revealed at a previous forum:

Based on vertical scenarios, precipitate and optimize a complete algorithmic engineering product system in the business.

Quantum Bit has learned that the "vertical domains" referred to here include one of the company's original tracks related to the medical and health care field. Infinite Light Years may launch industry-specific large models and related applications for the medical field.

The second domain is finance, e-commerce marketing, and other fields.

This positioning may consider training costs and computing resource requirements, as well as industry service capabilities. In a forum in Hangzhou in May, Qi Yuan mentioned that the cost of training a generalist is extremely high, and large models for industry applications may need to be trained to understand and reason at the level of "high school students" before focusing on one or several related professional fields.

Who is Qi Yuan?

Qi Yuan graduated from Huazhong University of Science and Technology with a bachelor's degree in automatic control. After obtaining a master's degree in pattern recognition from the Chinese Academy of Sciences, he worked on machine learning at the MIT Media Lab and obtained a doctoral and postdoctoral degree.

He first came into contact with AI in 1998, and his MIT thesis was related to Bayesian inference and probabilistic graphical models in machine learning.

In line with his early belief that "automatic control, pattern recognition, and machine learning are all branches of AI," Qi Yuan's research at the bachelor's, master's, and doctoral levels was all related to AI.

During his time at MIT, both Google and Facebook extended offers to him, but Qi Yuan ultimately chose to remain in academia. He later left MIT for Purdue University, where he served as an assistant professor in the departments of computer science and statistics, and in 2013, he obtained a tenured position as an associate professor.

Just a year later, Qi Yuan returned to China and joined Alibaba, witnessing the company's IPO a year later.

Qi Yuan once stated that this choice had both the significance of "learning for practical use," where technological achievements can be directly used by millions of users, and a sense of technical challenge.

Upon joining Alibaba, Qi Yuan, along with Zipei Tu and Rong Jin, co-founded the predecessor of DAMO Academy—iDST, and Qi Yuan was responsible for leading the research team in China.

Starting from the infrastructure, he led the construction of Alibaba Group's first large-scale distributed machine learning platform, which has now evolved into PAI 3.0.

In addition, he also led the establishment of Alibaba's first professional deep learning-based speech recognition team.

Later, he held positions such as vice president and chief data scientist at Ant Group, head of financial intelligence at DAMO Academy, and chief scientist at Ant AI.

In 2021, Qi Yuan returned to academia and joined Fudan University (it was revealed that this was due to personal and family reasons):

He was appointed as a distinguished professor at Fudan University, a doctoral supervisor, and the director of the AI Innovation and Industry Research Institute.

Now, Qi Yuan is once again embarking on entrepreneurship in the wave of large models.

The deeper reason for this may be glimpsed from Qi Yuan's lecture at Fudan University's School of Economics, titled "From ChatGPT to AI for Science."

In the lecture, he mentioned that based on the current development of ChatGPT, AI for Science will have longer-term development prospects in various fields, such as the combination of AI and large models in life sciences, meteorology, and fluid dynamics.

The content-intensive industries mentioned in passing, such as medical and health care, and finance, are precisely the landing fields targeted by Infinite Light Years' AI large models.

But the deeper reason may be that the tide of AI is rising once again.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。