Sina Finance News 2017 The 10th Golden Reward Creative Festival opened today in Shanghai, Yang Weiwei, General Manager of 360 Corporate Commercial Products Division, Vice President, Chief Scientist and Dean of Artificial Intelligence Institute, and Associate Professor of National University of Singapore, Yan Shuicheng, An Bu Si Pei China/North Asia Digital Performance Brand CEO, Chen Jianhao, Chief Innovation Officer of China Electronics Group, Dentsu Anjisi Group, Tian Gui Founder, CEO Xue Guirong, and Executive Deputy General Manager of Suning Yunshang Group Advertising Business Center Ding Haiyun attended the “Future has come†-- Scene thinking, creating 360° intelligent marketing" 360 special session and discussion.
The following is an excerpt from the discussion:
Yang Weiwei: In the era of AI, we want to see what impact AI will have on the future and what challenges the brand faces. Today, I have invited several masters in the field of artificial intelligence. Professor Yan, Dr. Xue, also invited the real experts in the brand world. Last year, Jin was invited to participate in our forum. We will do some research on this aspect.
Professor Yan, I know that in the past two years, artificial intelligence has been doing a lot of work for commercialization. Can you give me some specific introductions?
Yan Shuicheng: In fact, in 360 companies, early artificial intelligence is a comprehensive support for different business lines of the company. We are roughly divided into computer vision, one is conversational artificial intelligence, including voice and semantics, and the other is big data, supporting our IOT. Strategies, security strategies, short video strategies and other dimensions.
Specific to the direction of commercialization, we think so, in fact, commercialization from the perspective of artificial intelligence to do two things. One is to meet the individual needs of users, and the second is to satisfy the personalized experience of users. To meet the user's personalized needs is to say how to use the big data of 360, for example, from the search, security products to get a lot of user personalized data to establish user portraits, through user behavior to make click predictions better, This piece is very good for both 360 and other brother companies. Another aspect is that with the development of the mobile Internet and the development of artificial intelligence interactive technology, everyone is not satisfied with meeting the individual needs, and more hope that when the information of these advertisements and related products arrives at the users, they can be personalized. Experience. In other words, we hope that the unified graphic advertisements will not be presented directly to you. I hope that the way this advertisement is presented is related to the user. At this time, a lot of computer vision technology is needed. In the past few years, the technology of computer vision has developed very well, and it can put people on the face, for example, key points. Some AR technologies have reached the point of being productized and commercialized.
So I think that in the following period, 360 will make more use of visual related technology to make the personalized experience better in the process of commercialization. This is our general idea and is some of the points of commercialization of the 360 ​​Artificial Intelligence Institute.
Yang Weiwei: What I was impressed with is the individualized cognition. At the same time, we need a personalized experience. We have to bring a unique experience to each user. This gives us a big challenge. I have been saying before, I was talking yesterday. We said that the advertising platform, the advertising system, after the AI ​​came in today, we have a lot of work being replaced by AI. Because advertising companies are doing brand management and marketing services. What kind of challenges do you think will be brought to the advertising company and how to deal with it?
Chen Jianhao: Our theme is future is coming, in fact, future is now. Consumers on the website, recommended products, videos and advertisements, in fact, have the push behind the big data. For advertising companies or advertisers, this is a great opportunity for you to do better and more accurate delivery. Artificial intelligence gives the marketing field, we can sum up into three points. In the past when I was doing media placement, I wanted to optimize the media delivery, but all based on past experience. We now have artificial intelligence, but we can use one kind of Predictive form, artificial intelligence can transcend people, I think it can see the behaviors and patterns that people can't see, can see the people around you, the people you are related to, the behaviors produced in different time periods And then a new pattern will be generated. This is the best opportunity for advertisers to deliver.
Yang Weiwei: So not only do you have to guess what you like, but also guess that you will like it in the future. This is also a very interesting topic, and guess that you will be able to do solutions and recommendations in a more personalized way in the future.
Before his current startup, Dr. Xue was doing a very basic research on artificial intelligence. He was responsible for the basic platform of big data and display advertising in Ali, six years ago. When we talked yesterday, Dr. Xue told me that he had no interest in display advertising. I found it very interesting. I made an advertisement for 18 years. I still have a lot of confusion. I want to ask Dr. Xue to talk about it. You go out from this industry and stand outside to see the industry. Do you think this industry can talk about why it is not interested?
Xue Guirong: It’s not that I’m not interested. I think the challenge of this industry may actually be even bigger. When I was in Ali, I was responsible for the DMP platform. We do a lot of interaction with advertisers, we found that although you provide a good media, plus a good data quality system to give them, how many advertisers can use 100% of the system Get up? I feel that today, because of the ability I have seen, I may only use the value of 23% and 45%.
There are a lot of problems in this, many brands, many SME advertisers, if they want to use this platform well, they may need a lot of artificial intelligence technology. There are some technical things in it. I think the advertisers are actually quite technical. The knowledge of the advertisements learned by the advertised people is the latest. But when I came across such a big data resource today, I encountered such a good media. How much we can use, we can dig up the potential of this, which is what I thought a lot. It is Ali's slogan that makes the world difficult to do business, but the degree of difficulty in doing business should be that we can help the industry to be more efficient or more quality.
So, today there is just one chance. This thing comes out of AI, including 360 AI lab, including companies that are doing AI labs, including Ali's Dharma. In essence, we hope that through AI, brands or SME advertisers will have a better system to help them, and 100% of the value of the data will be fully utilized. This is actually a very challenging thing, because the amount of data, and very valuable, is that the amount of data is in this place, and the use for him is constantly increasing in value. Data is a good resource. Whether this resource can be mined the most, in fact, we must rely on us to do a good job.
We jumped out to see this thing today, such as whether we can make an AI system, so that each family has a very powerful system to help SMEs or brands better advertising. And from the perspective of ability and resources, it is possible to lower the price, instead of having an expert team with an AI. This may be the biggest challenge today. Can we do this, let AI help the brand or small and medium? Instead of recruiting a large team to do this, companies make efficient and valuable delivery. We hope that we can do this.
Yang Weiwei: So in fact, from a general model, go to a personalized model and go to a more real-time low-cost model.
Xue Guirong: Yes, for SMEs or brands, this model is no longer controlled by Ali, 360 or Baidu, but should allow advertisers to have better tools or better technology to support our pitchers. they. Probably this is the case.
Yang Weiwei: So in fact, you are not interested in the display advertisements of big companies and big platforms. You have to serve small and medium enterprises. Mr. Ding, we all know that Suning began to study a lot of personalized experiences long ago. Today, different people enter Suning and often see very different. Then we actually talked about this, in fact, for the e-commerce platform with a large number of commodities, in fact, can do a lot of things. How to ensure a consistent brand experience, brand promise, and brand image in the personalized experience, this is the doubt that has been haunting me.
Ding Haiyun: Thank you, Wei Wei, thank you all. From this point of view, including the views of the previous experts, talk about some of our ideas and practices.
In fact, from the development of brand building to marketing, to today, it should be said that the application of big data, including the continuous improvement of artificial intelligence technology, in fact, the help of the brand is obvious. It should be said that artificial intelligence promotes the evolution of the brand and the brand's reach between users. It enables brands to better understand users and insight into users. At the same time, in the application process of big data, we collect the user's current information through the user's behavior, including his age, gender, shopping needs, preferences, including his brand's cognition, loyalty and so on. In fact, after the whole big data analysis, through the intelligent advancement, we can deeper feedback to the brand level to understand, what is the status of our brand in the minds of users.
Therefore, through this kind of research, we now have more to do big data and intelligent things, and communication and product information, and reach a close contact with users. The purpose of this closed loop is to enable many marketing to plan ahead in the actual operation. Including the later decision-making process, a lot of work can be made clearer.
It should be said that from the beginning of the planning to the final decision, we hope that artificial intelligence can help us to automate the selection of resource media, including the delivery and distribution of automation. In this process, we all hope to have a more objective and more regular thing, so that we will have a lot of feedback and help for our later optimization.
Today, we are more worried or need to further recommend, in fact, how can we better study the psychological dynamics of users, you can use big data to do analysis, but now artificial intelligence should be better deduced to tomorrow, the day after tomorrow What are the needs of users in the future? This is the direction that artificial intelligence or big data applications give us guidance, or the effective direction. To help us better understand the user, it is also in the marketing behavior, let us predict the user's needs in advance, from the user experience, there will be a better experience. This is also the aspect of the combination of product and effect that marketing people have been working on. Thank you.
Yang Weiwei: I actually think that the experts of both brands mentioned a word and forecast. That is, I want to predict what your next needs are, and predict which consumers, besides the consumers we already know, may become our target consumers. I think this topic is very interesting. It is actually not in my category, but this topic is very interesting. The experts of the two brands have jointly proposed it. I want to ask an expert in artificial intelligence to talk about it.
Yan Shuicheng: The prediction is quite interesting. I am a commercial layman. After I contacted 360, I began to think about what kind of real problems might exist in commercialization. After a period of exchange, it is quite interesting to say that there are many factors in the current forecast, and we have not considered it.
We are more concerned with what a person's previous behavior is, and what the previous behavior of similar people is. I hope that after counting this information, I can predict what will happen later. In fact, there is another important factor. It is said that many people's needs and demands may be interfered by the outside world. There is no way to get information within the system. For example, now, there is an important thing to happen, then it is possible. This will make a big difference to our needs. But this sudden event is not included in the traditional statistical model or the general learning-based model.
How do you factor in factors that may affect your needs or demands? I think that if you can think more about it later, it should be very valuable. But this thing is more difficult to do, because some related data, it is in the hands of different relatively large companies, for example, the data of this sales is mainly on the e-commerce platform, for example, social data is On social platforms, there are some news related data that may be in another person. You can think that the data of different forms are related. As we seem, it seems to form an island of data. How to get through the data, establish their links, and collaborate on related factors, you can explore it, so that your future predictions will be very valuable. Sometimes it may not be possible to predict immediately, but there is a factor that will trigger another demand.
At least in the system I have seen, this factor has not been explored very well. It should be of great value if you find it well.
Yang Weiwei: Just like talking to He, I said that if the Chinese team is out, the sales of Chinese football will increase by more than 100%. This is triggered by an external factor. I know that Dr. Xue has made a game of Go.
Xue Guirong: This is a little bit related to this. We have also made a system similar to AlphaGo. We feel that it is very challenging to predict this. You can tell us about it. It's very simple. There is a problem in it. Let's go next. How many times do we have to go this one? We want to call the cluster of GPUs in the background. We want Calculated about 200,000 times. We make a decision to do 200,000 inference calculations, which is very time consuming. If you really want to be very extreme, you have to consider a lot of factors. We not only consider the human factors, but also the factors of the scene and the factors that interact with the real-time environment, which will affect our personalized forecast. Precision.
But today, in fact, the problem we have seen is that everyone’s predictions may have turned into a big company to help us filter the data. What do we think of this? Advertising hegemony. What is advertising hegemony? It is this matter that is not the decision of our advertising company or brand or SME advertisers. In fact, the decision is in the hands of a programmer. He thinks this algorithm should do this. Today is the case. But the future should not be like this. We talk about the consumer experience, my space is my master, this thing should be the direct communication between consumers and brands, may be better, have a good experience.
Today may be a platform-based display of advertising in the control of the media, but in the future we hope to have AI things, can make better predictions, more real-time predictions, or more accurate feedback predictions. We hope that this thing will become smoother and make a better experience.
Yang Weiwei: It seems that it is not so easy. There are many external factors in addition to the historical behavior outside the consumer's own scene. These will become very complicated after being put in. It seems that we still have a bowl of rice to eat.
Yan Shuicheng: I have spent a lot of money on calculations and spent a lot of money on it. Is there any extra space?
Yang Weiwei: From the perspective of a professional brand, under the premise of ensuring the user's personalized experience, for example, we have some small and medium-sized advertisers. Everyone knows that everyone is talking about information flow today. The information flow is basically a picture with a line of text. This text, we have prepared thousands of words for an advertiser, which is what the second yuan said. It is said to the technology otaku. The product recommendation is not to put the title of the product out, but also to do a lot of heavy Write to meet the user's personalized experience.
I still have such a question. How does the brand have a unified logic behind the personalized experience, and the unified brand runs through?
Chen Jianhao: Under the premise of personalization, AI is a process for giving marketers a way to see consumers and understand consumers. That is, he really understands the intention behind you, so through this clearer understanding of the requirements, even you can clearly understand your clear intentions from the emotional, from the expression pack, and even the typos that are entered. This brand has a good chance to pass on its own information. Just now, Ding also mentioned that it is necessary to consider the construction of the entire brand culture and the emotional resonance with consumers. Therefore, in this part, I feel that there are many opportunities for advertising companies and agents to have a good opportunity. How to give the brand spirit we want to pass to consumers in thousands of ways, and then generate the biggest The effect of the brand plus effect.
Yang Weiwei: So in fact, in a more closed system of information that seems to be more closed, how can the brand capture this person and resonate. This is a very good topic. The time is almost up. There are still many topics to talk to, thank you very much. Thank you.
Sina statement: All meeting records are on-the-spot shorthand, without the review of the speaker, Sina.com posted this article for the purpose of transmitting more information, does not mean agree with its views or confirm its description.
Enter [Sina Finance and Economics Unit] Discussion
Rubber grommets are essentially rubber molded parts which are placed around drilled or punched holes to an opening in a surface. Rubber grommets mechanically eliminate sharp edges providing a cushion of protection for the items which passes through the surface material from tearing or breaking wire, cable, tubing, piping, hoses, air lines, rope, etc, add aesthetic value to your applications. In addition, many grommets also serve as seals.
Custom Rubber Grommets ,Rubber Wire Grommets,Silicone Rubber Grommet,Colored Rubber Grommet
Xiamen The Answers Trade Co.,Ltd. , https://www.xmanswerssilicone.com