The following is the Four suggestions for successful application of machine learning in Enterprises From Amazon cloud technology Gu fan recommended by recordtrend.com. And this article belongs to the classification: Industry information.
Author: Gu fan, general manager of Amazon cloud technology Greater China Cloud Service Product Management
With the continuous development of machine learning, more and more traditional enterprises have begun to apply machine learning for business innovation and business remodeling. McKinsey’s previous special report on the impact of artificial intelligence on the world economy shows that artificial intelligence is expected to contribute 13 trillion US dollars to the global economy by 2030. The traditional fields such as retail, transportation, logistics, manufacturing and agriculture, after being enabled by artificial intelligence and machine learning, will drive far more economic contribution than software and information service industries. Moreover, compared with software and information services, these industries also need to be empowered to help them better deploy and use artificial intelligence and machine learning.
Combined with Amazon’s 20-year innovative practice of machine learning in the world and Amazon’s cloud technology’s experience in helping more than 100000 customers use machine learning in the cloud, we summarize four suggestions on the successful application of machine learning in traditional enterprises.
1、 Develop a clear data strategy
The three elements of machine learning include data, algorithm and computing power.
For most industries, collecting and processing data is a difficult problem. Therefore, before enterprises begin to apply machine learning, they need to fully understand their data status and formulate data strategies. What data is available now? What data can be made easy to use by a certain amount of work? If you have identified several scenarios that you want to try to use machine learning, you can use the reverse working method to extrapolate backward according to the goal, and calculate the required data, the data that you have and the data that you still lack, and the data that you need to collect from now on. Only by solving these problems and formulating a clear data strategy can we have real and well prepared data to meet the needs of business innovation based on machine learning.
Although some customers collect a lot of data, the data readiness is low and the data quality is not high. For example, if there are outliers or missing values in the data sent by the sensor, then the probability of the trained model is not accurate enough by applying machine learning on this basis.
2、 Starting from the right scene
Enterprises often have a myriad of ideas when they apply machine learning, so what projects should we start from to carry out machine learning? Here is a decision-making reference framework, which can be evaluated from three dimensions: data ready state, business impact and machine learning applicability. Enterprises can choose application scenarios with high data readiness, business value but low business impact and high applicability of machine learning as machine learning pilot and demonstration projects.
Specifically from three aspects. First, in the early stage of machine learning, the company may still have some doubts about its role. Therefore, we need to start with an innovation project with relatively small investment. It will not affect the company’s core business. Once it is successful, it can help enterprises accumulate experience and win internal trust.
Second, the project needs to have business value as well as be suitable for machine learning. Third, to find a scene, machine learning is only used as an auxiliary part to speed up the work of automation, not a substitute for human. For example, a doctor’s diagnosis of a patient is composed of many links. The process of looking at ECG and X-ray can be automated by using machine learning to speed up the doctor’s diagnosis process. However, machine learning will not replace the doctor’s work, which has less impact on the treatment process itself and is easier to get the doctor’s support and cooperation.
After the successful delivery of several small projects that can be completed in three to six months, the enterprise will have enough confidence and motivation to obtain the support of the leadership team, increase the investment in machine learning projects, and gradually apply machine learning to transform the core business.
Take harvest Wealth Management Co., Ltd. as an example, it is an independent wealth management organization under Harvest Fund. It has wealth management service centers in major cities in China, and will create more than 3.1 billion yuan of return for customers in 2020. As a financial enterprise, harvest wealth takes the media platform as the breakthrough point, and with the help of the standard AI capabilities provided by Amazon cloud technology, including out of the box tools and the AI model customized on the machine learning service platform, forms a media processing platform integrating media capital storage, voice transcription, short video generation and personalized recommendation, So that financial enterprises have the opportunity to accurately recommend the financial videos that customers need to reach customers from more channels.
3、 Professionalization of data scientists
The third suggestion for the successful application of machine learning in enterprises is to professionalize data scientists. Take Amazon’s construction of a machine learning team as an example. In Amazon, instead of putting data scientists in a central team, we put data scientists together with product and business teams to make data scientists business. Amazon is committed to customer-centric, and our machine learning scientists should start from improving customer experience, rather than studying machine learning algorithms.
The professionalization of data scientists is an important experience of Amazon. We copied this experience into Amazon cloud technology’s customer projects. When customers lack data scientists, Amazon cloud technology’s data scientists and engineers will join the project team to work with the customer’s business development team, gather the strength of data scientists and domain experts, and make innovations to improve customer experience.
Traditional enterprises usually do not have experts and data scientists who are proficient in both business and machine learning technology, so they can also put data scientists / machine learning technology experts together with business experts to achieve technological innovation. Shandong Zibo Thermal Power Group successfully twisted its business experts and machine learning technology experts into a rope through the empowerment of Amazon cloud technology. Zibo thermal uses the rich artificial intelligence and machine learning technology and services of Amazon cloud technology to jointly develop a set of intelligent heating expert system based on machine learning. According to the weather, SCADA industrial control data, building maintenance structure and other information, it calculates the best heating mode, and gives specific operation instructions to achieve accurate heating, It can not only keep the best thermal comfort temperature of human body at room temperature, but also save energy and increase efficiency.
4、 Coping with skill gap
At present, one of the biggest bottlenecks in the deployment and application of machine learning in most enterprises is the shortage of machine learning talents. Emerging companies compete for machine learning talents, while all kinds of traditional enterprises also need machine learning talents. In this case, the best way to solve this problem is to find a service provider who can help and empower the enterprise.
In the communication with industry customers, we found that many industry problems need a lot of iteration and optimization of the algorithm to continuously improve the accuracy. Some industry problems even need to study new algorithms to solve. In the face of these complex industry problems, we can’t just teach customers to use tools. Amazon cloud technology’s approach is to “help the horse, send a ride.”. We have gathered solution architects, artificial intelligence labs, data labs, rapid development teams and professional service teams to participate in the life cycle of the project according to the needs of customers. We work with customers to find business scenarios suitable for machine learning, develop product prototypes with business personnel and technical personnel, and then implement them in a rapid iteration. Let the customer complete the product prototype development with as little trial and error cost as possible to make up the skill gap of the customer.
In this way, we teach people to fish, enable customers to innovate, and always adhere to platform thinking, so that more people can use Amazon cloud technology to create and invent, and make artificial intelligence and machine learning inclusive. Amazon founder Jeff Bezos said that innovation has many forms and scales. The most radical and transformative innovation is to help others release their creativity and realize their dreams.
One of the core goals of Amazon cloud technology is to deliver machine learning capabilities to every developer. With the help of Amazon sagemaker’s ability to help customers quickly build, train and deploy machine learning models, we are able to further deliver machine learning capabilities to more users who want to innovate based on machine learning.
In a word, there is great potential for enterprise customers to apply machine learning. It is suggested that enterprises should formulate clear data strategies, find suitable application scenarios for machine learning as the entry point, first break through the innovative business, and then transform the core business. At the same time, let the data scientists go deep into the business and avoid making cars behind closed doors. Hope that more and more enterprises through machine learning to achieve continuous innovation and development, in the fierce competition in an invincible position. Read more: AWS releases Amazon Devops guru, a new machine learning driven operation service. Amazon cloud service (AWS) expands the circle of machine learning and touches every AI worker. Amazon cloud technology launches predictive maintenance service for industrial equipment based on machine learning. Amazon cloud service (AWS) is officially launched in Ningxia and Beijing of China Sagemaker AWS Zhang Xia: 85% of tensorflow’s global load is on AWS platform, and the development cost can be reduced by 54% big data and AI strategy – investment oriented machine learning and alternative data methods (with 280 pages report) the impact of machine learning on SaaS industry Amazon cloud service (AWS) accelerates the landing of cloud products and services in China: it takes only 10 seconds for AWS to analyze users’ interests with machine learning. AWS publishes five machine learning services for industry, from professional field to mass field. AWS joins hands with China’s local travel giant FAW car Hailing Amazon cloud service (AWS) to promote machine learning in an all-round way How does Xi Xin apply machine learning to predict traitors in the game of rights
If you want to get the full report, you can contact us by leaving us the comment. If you think the information here might be helpful to others, please actively share it. If you want others to see your attitude towards this report, please actively comment and discuss it. Please stay tuned to us, we will keep updating as much as possible to record future development trends.
RecordTrend.com is a website that focuses on future technologies, markets and user trends. We are responsible for collecting the latest research data, authority data, industry research and analysis reports. We are committed to becoming a data and report sharing platform for professionals and decision makers. We look forward to working with you to record the development trends of today’s economy, technology, industrial chain and business model.Welcome to follow, comment and bookmark us, and hope to share the future with you, and look forward to your success with our help.
-
-
-
-
