Netease Technology News December 8 news, 2018 NetEase economist annual meeting artificial intelligence forum, SF chief scientist Liu Zhixin delivered a speech.
Does artificial intelligence have the potential to replace humans? Liu Zhixin believes that artificial intelligence is to help rather than replace human work.
In the logistics field, for example, everyone thinks of a courier, but through appearance, logistics is a complex supply chain, accompanied by a complex network topology, which requires many complicated algorithm models to solve. Artificial intelligence can help optimize operations, enhance complex decisions, and enhance learning methods to help make decisions more efficient. These jobs and human work are complementary and do not replace human work.
In addition, logistics naturally requires a lot of manpower. Artificial intelligence frees people from simple and repetitive tasks and allows us to do more complex and valuable work. Help customers get a better experience.
Liu Zhixin used three examples to explain how SF has used artificial intelligence to assist work.
The first example is the digitized brother created by AI Assistant, 200 million times of sorting per day, 600 million operations, very heavy asset method, and valuable resource acquisition methods that are difficult for other companies to mine. The sixth generation of handheld terminals can complete the work of collecting data. Afterwards, he developed a smart bracelet to liberate his brother's hands. For heavier cargoes, Shunfeng has developed a robotic arm. In addition, AI can help optimize the route.
The second example is the intelligent customer service with temperature. The vast majority of customer service is carrying out simple and repetitive tasks, automatically assisting customer service personnel in operating and allowing customer service personnel to perform more temperature-based services. The third example is the assistance of AI in customer management and junior management.
The following is the full text of Liu Zhixin's speech, slightly edited.
Liu Zhixin: Ladies and gentlemen, good morning everyone!
Thank you very much for hosting the invitation. Let us have this opportunity to share with you some of the work we have done in SF, especially in the logistics industry. This is the effect and changes that artificial intelligence brings to this industry.
Today's theme is "Artificial Intelligence and the Future of Humanity." I think everyone is more concerned about whether artificial intelligence will replace the possibility of our human work in the future. I will share with you some of the work we have done in the area of ​​artificial intelligence, exchange our understanding of this issue, and see how artificial intelligence can help rather than replace human work.
First of all, let's look at what the relationship between artificial posts in the logistics field and artificial intelligence is. Speaking of the logistics and express delivery industry, everyone may be the first to imagine the screen is the endless stream of small brothers, and on the road to use the truck, including the airport's freighter and so on. Through these representations, we actually see the following is a very complex long chain of business links, including small brother pickup, to the sorting and then to the destination, and then a sorting, further refinement, do some Branch transportation, as well as the final courier and so on. This business chain is very long, and it is accompanied by a very complex network topology. How to make massive packages efficient and fast in the middle of a very responsible topological network structure, and to complete its journey to the final destination at a low cost, that is, we have a lot of statistics, operations and other algorithms and mathematical models that need to be solved. a problem.
The rapid development of artificial intelligence today can also help us improve the capabilities and efficiency of the original model. For example, when we were making predictions, the traditional method was to use time series. The development of machine learning now allows us to combine multi-dimensional internal and external data to build multi-dimensional models of machine learning and train him to improve the accuracy of our predictions.
As another example, when we are doing route planning, we will traditionally use a lot of prediction and optimization methods. But when our volume is very large and the situation is complicated, we need to make some complicated decisions. At this time, we will use reinforcement learning methods to help our decisions be more efficient and smarter. We can see that the development of artificial intelligence technology actually helps our original algorithm to be more efficient, and this algorithm itself is complementary to artificial posts. It is a tool and will not replace people's work.
From another perspective, the logistics field is a direction of high demand and high traffic. Naturally, we need to have a large number of jobs. Returning to the nature of artificial intelligence, artificial intelligence means that we use data to do four things: first, perception; second, cognition; third, modeling; and fourth, planning. What we do is these levels of things, and more to assist humans to complete their work more efficiently. So we research AI and research big data. The fundamental purpose is to make our employees work more efficiently and free them from many simple and repetitive tasks to accomplish more and more valuable, more challenging and more Some of the hard work brings a better experience to our customers.
Based on the two points just mentioned, we can see that China's logistics industry has now reached a very challenging, but it is also a very exciting period. As a leading company in the logistics express delivery industry, we have the responsibility to guide the future development model of China Logistics.
Below we can share with you a few specific examples. How our artificial intelligence algorithm helps our staff to complete their work more efficiently and allow our customers to have a better experience than to threaten existing manual work. Posts.
In the first example, AI helped build a digital younger brother.
Every day we have tens of millions of parcels, about 40 million kilograms in weight. After our dozens of first-line brothers, after 200 million sorting and 600 million handling operations, these parcels can be sent to the whole country. Various places. How do you think about this issue? Many people may think this is a very heavy asset industry. Our view is not limited to this. It is not only a heavy asset, but also a very valuable resource for data acquisition. Because our hundreds of thousands of little brothers are spread across all the communities, each office building is like Capillaries, collecting a large amount of data close to the physical world. His judgments and decisions are very valuable, and they are also very difficult for other companies to obtain. So, we must fully tap the value of the data that these little brothers bring to us, because the data itself is an important engine for the direction of artificial intelligence.
In the last year, there was a fifth-generation handheld terminal for the younger brother. Its biggest function was scanning and operation. However, we have found that it does not help us to complete the collection of Xiaoge's daily behavioral data. So we quickly invested resources to develop the sixth generation of handheld terminals, more to complete the work of collecting data and so on. We are also faced with the problem that our younger brother usually needs hands to work during his daily work. If he has a handheld terminal to carry out some work, it will be very inconvenient. Therefore, we have developed a smart bracelet that can help him more. Efficient and convenient work. In addition to the ecological end, we also have a lot of work for decoration, sorting, loading and unloading. One is the work intensity is very large, and the other if it is a very heavy item, in fact, there is a high risk. So we also invested resources to develop the robotic arm to help Xiaoge to complete his work more easily.
Back to circuit planning, circuit planning is a very central issue in the field of logistics. We have combined traditional methods to do some work in this area, optimizing our routes and decisions under many complex conditions based on real-time quantities, making our logistics more efficient and more cost-effective. This is mainly for the example of AI assisting express delivery.
The second example, "intellectual" creates temperature customer service.
How do you think about smart customer service? We want to provide customers with a more personal, more customized, and more temperature-based service. However, we actually found that most of our daily work of our customer service staff is repeating simple, monotonous work. Based on this problem, we have adopted some NLP technologies and some natural language processing techniques to analyze the intentions of the customer dialogue. After extracting his key information, it will help our system to automatically assist our customer service staff or automatically. The completion of some operations, so that customer service staff really put their time, ability and resources into a more customized and more temperature-sensitive scene to the customer, truly complete a person-to-person exchange.
In the third example, intelligent decision-making aided in manual decision-making, helping the company to automate, informatize, and intelligently and efficiently manage the development. .
After escaping from the specific first-line business, the same SF Express will also face many management challenges. Traditionally, management is based on human perception, experience, or some predictive rules. This time will be limited by the individual's ability, but also by the fact that we do a lot of time is a partial judgment, not a global judgment. So, we have also combined very internal and external data from many dimensions to create some machine learning models that help us make some smart decisions.
Here are a few specific examples:
The first is the management and planning of the terminal field.
The terminal field is a very important part where a large number of goods enter and leave for sorting. At this time, we used computer vision technology to help us build some models, simulate the entire operation, and help our operations in all aspects of the terminal field more efficient, making our resources more flexible, controllable, and more flexible.
The second is about the management of Xiaoge.
This is the aspect of assignment tasks. We will discover through actual data that different dispatch tasks are actually very different, such as areas, etc., and even include when the user is home, etc. At this time, we will According to the data, we can find out some different rules for each task, and at the same time, mine what the little brother is really good at doing, and match this task with the ability of the little brother. At the same time, we can also achieve complementarity and make resources more flexible.
The third is about customer management.
Traditionally, we will have a dedicated and relevant team to manage all our small and medium customers and large customers. For example, at this time, they will regularly analyze these customers every month whether they will have some changes, whether there will be some risk of loss. But this has a very serious lag, and I can't do these analyses on a large number of customers every day. Therefore, we are also from a large number of data to dig out some of the laws of the customer, will do some machine learning, we have traditionally defined many rules, and this rule is not covered. Now we can do real-time, even based on customer's data such as quantity and order, we can predict whether he has the risk of losing, and allow our sales staff to intervene the first time, communicate with customers, help them reduce losses .
From these several examples, we can see that artificial intelligence itself can help our first-line little brother staff, can help our second-line customer service staff, and can also help our third-line managers. In general, what is more complete is how I use data and algorithms to help us to be more efficient, more effective, and to complete some of his work more easily than to replace people in this area.
My speech is here, thank you!
Helmet speaker driver used in the helmet, which can delivery vocal sound and music for the person who die the bicycle,cars, motor car, motor car and so on. Helmet Speaker,40Mm Helmet Speaker,32Ohm Helmet Speaker,Motorcycle Helmet Speaker. our XDEC Speaker is a famous speaker manufacturer who doing these products, for now we have 30mm, 40mm 50mm 57mm headphone speaker driver, for you, choose , welcome contact us for more choices.
Helmet Speaker,40Mm Helmet Speaker,32Ohm Helmet Speaker,Motorcycle Helmet Speaker
Shenzhen Xuanda Electronics Co., Ltd. , https://www.xdecspeaker.com