The AI development prior to 2017 focused more on the technology itself. Startups with founders who can build great technologies or who came from famous tech firms were relatively easy to win the trust of investors. 4paradigm was such an AI startup with well-known founders.
As a startup founded in May 2015, 4Paradigm quickly won the trust of Sequoia Capital China and obtained an Angel Round financing of CNY 5 million (estimated by EqualOcean), with a valuation of CNY 40 million. After the Series A & B financing in 2016 and 2017, 4Paradigm’s valuation reached more than CNY 600 million (USD 100 million).
The smooth funding process relied heavily on 4Paradigm's two founders, DAI Wenyuan (戴文渊) and YANG Qiang (杨强). DAI Wenyuan, the Founder and the CEO, is a leading scholar of Transfer Learning (a branch of deep learning), the World Champion of ACM-ICPC, the youngest Baidu senior scientist in history, and a former director of Huawei Noah's Ark Lab. DAI has directed Baidu's online marketing system, Baidu brain, as well as other core products, and enabled Baidu to grow its revenues by 8 times in 4 years, a rate consistently beat Wall Street Analysts' expectations.
YANG Qiang, the Co-founder and Chief Scientist who used to be DAI Wenyuan’s supervisor in The Hong Kong University of Science and Technology, is the first Chinese academician and the only Chinese executive committee member of international academic intelligence association (AAAI), the Editor-in-Chief of IEEE Big Data Journal, the only Chinese Chairman of the Board of Directors of IJCAI, ACM Distinguished Scientist, the champion of KDD Cup, 2017SIGKDD Outstanding Contribution Award winner. Furthermore, as the dean of the computer science department in HKUST, he has published more than 400 paper which was cited for more than 20,000 times.
Besides the funding rounds, 4Paradigm’s technological focuses are also determined by those two founders. Transfer Learning, the main research direction of DAI and YANG, is a research topic in machine learning which focuses on storing knowledge and applying a pre-trained model to another problem. For example, the knowledge gained while learning to recognize cars could apply to the case of recognizing trucks (Description from Wikipedia). 4Paradigm claimed Transfer Learning as the next generation technology of AI in the industry, and with such technology, DAI Wenyuan hoped to make AI for everyone.
In an interview with EqualOcean in February 2017, DAI Wenyuan shared the positioning strategy of 4Paradigm, which is also related with the “AI for everyone” prospect, to provide general technology of artificial intelligence to different industries, as a platform.
“What artificial intelligence really creates is a large amount of labor force for the society. The value of 4Paradigm would not be a better algorithm of face identification but helping others to make better ones. So far, a unprofessional staff could use our platform to achieve professional AI output, with trainings of two to three weeks. I believe AI will definitely be a much bigger business opportunity than the Internet” said DAI Wenyuan.
In order to make everyone, every company could use AI, and afford AI as well, 4Paradigm introduced the company’s flagship product – the AI Prophet, which was initially launched in 2015 and upgraded to its 3.0 version in late 2017, is an AI development platform that enables enterprises to effortlessly build their own AI applications, and thereby significantly increase their operation’s efficiency. Beyond this, it is said that 4Paradigm has also successfully developed more than 100 AI solutions for use in various settings such as finance, telecommunication and Internet applications. These solutions include smart pricing, real-time anti-fraud systems, precision marketing, personalized recommendation and more. However, 4Paradigm never published the sales unit or revenue of this product.
In 2018, about the evaluation of the capability of an AI company, the industries considered more and more of the actual income of a company. Besides the dream of “AI for everyone”, 4Paradigm had to pay more attention to the company’s cash flow, and so far finance would be a great industry to obtain revenue, because the industry is more data-intensive, and need better risk management, AI could do a much better job than human beings on the massive data processing. Thus, besides the platform positioning, 4Paradigm also tries to provide AI services to banks, as a Fin-tech supplier.
While on the other hand, finance in China is also an industry under strong regulatory, policies may have huge influences on the businesses. A possible trend is, bank systems’ usage of outsourcing suppliers would be further limited and even forbidden, and if this comes true, it would be a disaster for the bank suppliers.
For better serving the banks and ensuring the qualification of serving, 4Paradigm obtained its series B+ financing from ICBC, CCB, and BOC, three of the biggest government-owned bank of China. But what if the policies become stricter that banks can only use their own technologies? Does that mean those companies that cannot give up their financial businesses have to choose the path of further acquisitions by the banks? It is still unclear right now.
One way to the dream, difficult while beautiful; one way to the survival, with cash flow as well as risks to be swallowed, how to choose? This will be a “The Moon and Sixpence” question for 4Parandigm and its core team.
- Author: ZHANG Fan; LIU Huan contributed to this article. Write to ZHANG Fan at ZhangFan@EqualOcean.com