【2020 Solutions】 Lean AI, open source intelligent manufacturing helps companies quickly build AI teams
Does AI cost a lot? Is importing AI time-consuming and labor-intensive? How to build consensus within the enterprise and build a solid AI team? All the above problems are common problems faced by enterprises in AI digital transformation. Huang Mingshi, the founder and CEO of the AI startup Kaiyuan Intelligent Manufacturing Company, provides customized AI solutions on a "subscription basis" so that traditional enterprises that want to undergo digital transformation can quickly introduce AI solutions.
The term "Lean Production" appeared for the first time in the book "The Machine That Changed the World" in 1990. When it comes to business operations, there is no waste of resources. Phenomenon, the process operates smoothly and creates the most profit with the minimum investment.
AI subscription service solution to quickly assist in importing tools
Huang Mingshi graduated from Jiaotong University and received a PhD in Electrical Engineering from Penn State University in the United States. He also worked for a start-up company in Silicon Valley for five years. He was the chief data scientist and led a team of 10 people to develop multiple AI projects, including real-time image recognition. , 5G+AI, prediction system for dynamic expansion of cloud resources, etc. Huang Mingshi returned to Taiwan to start his own business and established Open Source Intelligent Manufacturing in May 2019. The development direction is to promote AI subscription services. He hopes to make AI practical and practical, help small and medium-sized enterprises with AI application needs, and accelerate the cultivation of practical capabilities. of AI talents.
Huang Mingshi believes that the scale of AI project software and hardware equipment, which can easily cost NT$2 to 3 million, is an "unbearable burden" for small and medium-sized enterprises with limited funds and human resources. However, for those who lack For small and medium-sized enterprises with limited resources, AI can indeed solve the problem of automation, reduce costs and improve efficiency for enterprises in a short period of time, and is an indispensable tool for digital transformation.
Therefore, Huang Mingshi follows the principle of "lean production" to prepare AI digital transformation tools for small and medium-sized enterprises, using small projects to get started, from corporate health clinics to identify problems, corporate training, consulting to providing AI model solutions. Done within a year. In the process of project promotion, we assist the company's middle- and high-level managers to empower them and help them understand the company's pain points, what problems AI can help solve, and introduce benefit analysis. By then, it will be relatively easy to introduce medium and large-scale AI projects.
Open Source Intelligent Manufacturing’s customized solution for the legal industry is the development of a “food advertising text recognition and analysis” tool. A medium-sized law firm wants to help clients solve the problem of "identifying food advertising violations." According to statistics, there are more than 4,000 cases of advertising violations in the food industry every month. The traditional method is to assign 2-3 lawyers to search for exaggerations or claims of efficacy in food advertisements published by major media. The cost of each lawyer is calculated at NT$5,000-10,000, which means the overall cost is considerable. However, by collecting relevant information through the crawler system, and then importing AI technologies such as natural algorithm (NLP) to develop food advertising text recognition and analysis tools, the recognition rate can reach 90%, which can also greatly reduce personnel costs.
In addition, Kaiyuan Intelligent Manufacturing has successfully applied face recognition to applications such as smart tourism and smart door locks, achieving an accuracy of 95%. It has also used graphic recognition to help digital advertisers achieve the function of AI image removal. , reducing the time designers spend on repetitive memorization by more than 80%.
It is worth mentioning that for designers, it often took 2 hours to memorize photos in the past. Using the AI model, 1,000 photos can be memorized in 10 seconds, which is amazingly efficient. This means that designers do not need to spend too much time memorizing photos, and can use their time to come up with creative ideas. When photos are needed, they can use AI technology to quickly memorize them; for ordinary consumers, when making presentations ( When designing PPT, you can also use photos that have been reversed to speed up the presentation production time. In the future, it will also be connected to Google Flickr's personal photo album or image search, so that you can directly memorize the required photos and complete the task in one go.
At this stage, the open source intelligent manufacturing project is cooperating with APP manufacturers to remove the photos and create a free website to benefit more people working in the design industry.
▲Open Source Intelligent Manufacturing develops an AI model for hair back removal, the effect is comparable to that of professional designers
In the medical industry, Kaiyuan Intelligent Manufacturing has also developed obstetrics and gynecology organ image recognition technology and conducted education and training in schools to help students make correct judgments.
Kaiyuan Intelligent Manufacturing also cooperates with the Taiwan Suicide Prevention and Control Association to use AI models to find people who are emotionally distressed, have depression, or have negative emotions and comments on online forums such as PTT and social networking sites such as Facebook, and design Prepare a suicide risk assessment and submit relevant information to the Taiwan Society for Suicide Prevention and Control to prevent it before it happens and reduce possible tragedies.
Four-step import method to complete the goal within one year
In order to help small and medium-sized enterprises achieve the goal of AI application through subscription services, Kaiyuan Intelligent Manufacturing hopes to use methods to find enterprise pain points in the shortest time and at the same time enhance the commercial value of AI applications. The methods are as follows:
1. AI Discovery Workshop: Guide companies to explore their needs through workshops or corporate health clinics.
2. Enterprise AI training: The introduction of AI requires the full support of the company's senior managers and the consensus of all employees to be successful. Through a one-month training, it can quickly transform into an AI-empowered enterprise.
3. AI consulting: What is important is the technical feasibility assessment. Not a set of AI solutions can solve any problem. Different AI models must be established according to the different needs of the enterprise in order to prescribe the right medicine.
4. Subscribe to AI solutions.