4Paradigm, USD 2 billion dollar Chinese Unicorn aims to bring AI to every industry by turning managers into AI experts in days
Even after snapping another round of US$230 million funding from the big tech names Cisco Systems, Lenovo and top venture firm Sequoia Capital, the world still knows very little about this US$2 billion dollar AI company called 4Paradigm (in Chinese called Disi FanShi). Led by their top AI scientist Founder & CEO Dai Wen Yuan, the company is building an entire ecosystem to enable every business and organization in the world to adopt bleeding edge AI technologies, without ever hiring a single data scientist.
The company’s key milestone are well documented on Baidu Baike (China’s Wikipedia):
2014 | Company Founded in Beijing China by AI Scientists
2016 | A+ Round Funding from Sinovation Ventures and Sequoia Capital
July 2016 | Launched AI Development Platform PROPHET
Nov 2016 | Created AI Lab with China Everbright Bank
Dec 2016 | First Company in history to receive WuWenJun AI Science Award
May 2017 | Selected as Gartner 2017 Cool Vendor, as the first Chinese company with a PaSS business model in AI
Jun 2017 | Champion in QianHaiZhengXin’s First Transfer Learning Competition
Dec 2017 | World Champion in BNP Paribus’s International Hackathon
2018 | B+ Round Financing from Industrial and Commercial Bank of China, Bank of China and China Construction Bank
Aug 2018 | Launched Chronic Disease Prediction Product with Rui Jin Hospital Group
Sep 2018 | Created AI Lab with Intel
Sep 2018 | Launched AutoML and AutoCV products.
Nov 2018 | Launched Advanced Recommendation System for China Daily
Dec 2019| Closed C Round Financing from all Big Five Banks in China.
Feb 2019 | Selected as CB Insights Top 100 AI Companies
May 2019 | Broke ICDAR2017 Multilanguage Text Inspection World Record
Jun 2019 | Launched Enterprise Hardware+Software Integrated System SageOne
Aug 2019 | Created Smart Retail Joint Venture with YongHui Superstores
Apr 2020 | Closed C+ Round Financing with US$ 2billion valuation
WHY THE NAME 4PARADIGM?
The term “4th Paradigm” was first coined by Computer Scientist Dr. Jim Gray’s paper “The Fourth Paradigm: Data-Intensive Scientific Discovery” in 2005. Dr. Gray received the Turing Award for his contributions in the field in 1998. The evolution of science today is entering the 4th paradigm since human history:
Paradigm 1: Empirical Evidence
The ability of humans to observe patterns in nature through our basic senses and to conduct subsequent experiment to achieve similar results.
Example: Discovery of fire from putting two rocks into friction about 2 million years ago.
Paradigm 2: Scientific Theory
The explanation of an aspect of the natural world that can be repeatedly tested and verified using scientific method containing observation, measurement, and evaluation of results.
Example: Newton’s laws of motion, which can serve as an approximation to special relativity at velocities that are small relative to the speed of light.
Paradigm 3: Computational Science
The use of computers to run mathematical formula based on theories created by humans.
Example: Weather Forecast that takes into account a wide range of historical and real-time climatic variables.
Paradigm 4: Data Science
Discovery of patterns by computers themselves among vast data sets.
Example: Autonomous vehicle can gather and process images in the environment and accurately determine the road conditions to drive itself better than humans can do.
When Founder and CEO Dai Wen Yuan was looking for a name of the company, this term kept coming back to his mind. He decided that his vision for the company is exactly within this historical leap into a completely new era of science. He wanted to create artificial intelligence in our modern computers to to help everyone make better decisions.
REAL IMPACT OF AI
Human brains are designed to focus on a few most important elements and we typically make our decisions based on those priorities. In contrast, computers can process and analyze an infinitely larger amount of data compared to humans. This unique capability brings the ultimate benefit of AI, which is to help optimize our decisions in every aspects of our lives, by uncovering millions of patterns that we humans simply do not see.
Predicting Diabetes
Previously doctors at Rui Jin Hospital concluded there are up to 30 rules based on all the medical data if predict if a person will become diabetic in the next 3 years. 4Paradigm was able to develop the prediction technology by digging into the medical data of 200,000 diabetes patients, and identified a total of 500,000 rules that could contribute to the formation of diabetes. This new prediction model turned out to be 2–3 times more accurate than the previous method (without using AI). This prediction model was found to be 7–16% more accurate when compared to international standards including American Diabetes Association (ADA), Finnish Diabetes Association (FDA), and Chinese Medical Association (CMA). In August 2018, former U.K. Prime Minister Theresa May also visited the project at Rui Jin Hospital.
Rui Jin Hospital and 4Paradigm later released this technology to every one for free with an APP. Anyone can access the APP by simply scanning a QR code. Today every one now has the ability to be alarmed of their probability of getting diabetes in advance so that they can adjust their lifestyle (diet, exercise, etc.) accordingly.
While the best technology always seems so easy and seamless, the story of how they are created often was not told. When the first doctor at the Rui Jin Hospital reached out to the team at 4Paradigm in 2018, the company was not familiar with the application of AI in the medical field. But both sides believed there was a tremendous amount of opportunity in this untapped territory. The project team quickly moved into a small attic space inside a hotpot restaurant in order to work days and nights with the doctors and nurses across shifts. Their full immersion allowed the project team to quickly ramped up their knowledge of the smallest details in the hospital operations, and be able to achieve the results within only a few months. During the press conference of the result announcement, the cofounders at 4Paradigm laughed at themselves saying that they spent so much time working in that apartment/ office that they would feel quite sick and throwing up when tens of gigantic pots were all running downstairs during dinner hours.
Fraud Detection
To financial institutions like banks and insurance companies, fraud has always been one of the biggest threats. 4Paradigm was able to use their technologies to help one of the biggest Insurance companies in China to create a whopping 1.5 BILLION RULES to detect fraud, compared to the mere 1,000 rules previously created by their team of 200 top in-house data scientists. This new capability allows the insurance company to detect and stop 5 times as many frauds. This translates to multibillion dollar savings a year as well as priceless enhanced protection to insurance company’s client assets.
Letting AI be the Boss
The last 20 years of mobile technology development has seen most businesses in the world connecting to their customers digitally in some form of app or website. This is the automation of only the FRONT end of the business, meaning that companies now spend less effort and money on finding and serving their customers. The real opportunities that AI technologies can bring lie just behind that digital storefront, in the operations.
Traditionally, there are 3 layers of employees in any company:
Top Executives | responsible for high level strategic decisions. Examples: Major investments, Choosing Markets, Stakeholder Interests, Risk Management.
Mid level Manager | responsible for operational decisions. Examples: Sales policies, Marketing Campaigns, Product Pricing, Logistics fulfilment, Customer Service.
Front line staff | responsible for direct relations with end customers. Examples: Sales Team and Customer Service Representative.
The fact is that majority amount of day-to-day business decisions happen with the Mid Level Managers. These decisions of what products to show to which customers on the website, what inventory to restock based on their sales forecast are the crucial decisions that have to be done right. However today many of them still rely on their gut-feel or overly simplified excel tables. Simply put our managers at best are still living in the BI (Business Intelligence) world, thinking they can make the best calls by religiously focusing on the KPI (Key-Performing Metrics).
4Paradigm intentionally positions themselves as the Change Agent for bringing companies from BI (Business Intelligence) to AI (Artificial Intelligence). BI means you have decided that a few metrics are the most important to your business and you will make all your decisions based on them. For example, you track the TOP 10 GROSSING products and make sure you show them in the recommendation page because “most people buy them”.
On the other hand, AI means that you completely let ALL YOUR DATA dictate how you run your business. In our example, you would build a Recommendation Model that always show the MOST OPTIMAL products to each of your customer. And the way to decide on what to show is based on the applying the 100,000+ rules each one customer based on what they bought, when they bought, what they browsed, what is happening in where the live, whether their last shipping has arrived, and so on.
When you go from BI to AI, your revenue can only go up, sometimes way up! This is because you are serving every customer differently and providing what is ONLY RELEVANT to them. Imagine your bank now has assigned a dedicated Account Mangaer to you instead of you having to call the the Customer Hotline number and be put on the queue. The improvement in your experience of course will be massively better! Now think about Google, Amazon and Netflix. We are so used to finding things that we want or like with very minimal scrolling but we may not know how hard the machine is working behind the scene that creates that magical experience for us, every single time. Every company has loads of data about you, but only the smart one knows how to run AI on those data to bring you delightful experiences.
Addressing Talent Bottleneck
Of course, the Big Tech Giants with billions of dollars can hire the best data scientists in the world to create the powerful AI tools. What about the 99.9% companies out there that do not have the talents? That is the precise problem 4Paradigm is trying to solve with their AutoML platform. By building a PaaS platform that provide all the infrastructure to create individual AI applications, 4Paradigm is effectively giving the chance to every programmer to become a data scientist within days. It is similar to what Amazon has done with AWS, which took away the most difficult tasks and allowing a person with no coding experience to deploy an application to the cloud with ease. 4Paradigm believes AI should be easy-to-learn, easy-to-use, affordable and accessible to all. They launched the 4Paradigm Academy as a free platform for anyone to become AI literate (the service is currently in Chinese only).
Within only a few years, 4Paradigm has already achieved the highest market share in AI applications in China, among sectors of Finance, Retail, Energy, Government, Healthcare, Internet. But they are only at the beginning of their journey. The latest around of US$230 funding will be used to expand their services and footprint into the global markets. From the company website, we noted that besides from their Beijing (headquarter), Shanghai, Wuhan, and Shenzhen offices, they have also recently entered Singapore.
The AI Pack
Global Market Intelligence Firm IDC concluded in their China Artificial Intelligence Software and Applications Tracker Report (2019 H2) that “4Paradigm is far ahead in market share”, leading all other machine learning platform players in China, including Alibaba, Baidu, Huawei, and Tencent.
Internationally, iResearch releases its “China Artificial Intelligence Industry Research Report” listing 4Paradigm as one of only six companies comprising the machine learning layer, alongside Google, Baidu, Facebook, Amazon and Microsoft.
Visit ChinaPotion Research Store.
