I am familiar with big data

Image source @Visual China Article丨Tanker Tanker, Author丨Wang Ying, Editor丨Dan I am familiar with big data, what should I do? “My friend and I bought the same necklace together, and when I had to place an order together, I realized that the price shown on her was actually 30 yuan cheaper than mine!” “When buying fruit online, my colleague showed pitaya 0.99 yuan, but at the same time mine showed 6.9 yuan. Asking the customer service first said that the purchase was restricted, and the colleague was a newcomer, but the official did not explain it clearly. My complaint was also given as evidence. Not fully cancelled.” “I bought 20 masks on Taobao, and the same individually wrapped 20 pieces. I showed 68 yuan, and my friend showed 98 yuan. The price gap is big.” Picture / Black Cat Complaint This is not the first time that the topic of “big data kills familiarity” has caused controversy. The most complained are e-commerce, food delivery and taxi rides. Regarding the black cat complaint, there are still complaints submitted by netizens in January 2019 that have not been processed. In March of this year, Professor Sun Jinyun of Fudan University spent nearly 50,000 yuan on 800 taxi rides. The news of the real hammering of big data once again pushed this topic to the forefront. Figure / Weibo voting After netizens expressed dissatisfaction and condemnation of the phenomenon of “big data acquaintance”, there was finally an official response. On April 8, 10 Internet platforms including Vipshop, JD.com, Meituan, Ele.me, Daily Youxian, Hema Xiansheng, Ctrip, Qunar.com, Ruqi Travel and Didi Travel signed a letter of commitment, saying: ” No price fraud, no use of data advantages to kill them. “Our recommendation algorithm is different from other content platforms. We recommend the best travel route and the most cost-effective way according to different people’s travel methods. The recommendation system is constantly iterative, and it is impossible to solve all problems at once.” The big data algorithm engineer in the field told the explorer “Tanker”. So, how do these algorithms integrate into our daily lives step by step? What are the advantages and disadvantages of the “recommendation algorithm” for businesses and users? As Internet users, how can we protect our data privacy to the utmost extent? 1. Algorithms leave you invisible How do companies use data algorithms to give you “a clear view”? “(Enterprises) use algorithms and behaviors and guess what users are secret, not clearly stated by the platform, such as shopping interests and other behavioral interests. The concept of guessing a person’s interest or shopping tendency through behavior is not new. People will guess the information through the conversation with the other party and what they wear. They call this “Bahuang”.” Liu Peng, the author of “Computational Advertising”, vividly explained the basic logic of the recommendation algorithm in the popular science video, his other identity It is Zhihu Big V “Bei Ming Cheng Hai Sheng”, he was in charge of big data monetization related business in a technology company. In addition to the appearance of professionals in the science and technology circle, the scientific research team in the academic circle has further verified the “magic calculation” of the recommended algorithm. In 2020, Professor Sun Jinyun led a team of more than 20 people to conduct more than 800 field surveys in Beijing, Shanghai, Shenzhen, Chengdu and Chongqing. He spent nearly 50,000 yuan in taxi fare and reached an amazing conclusion, that is, the user’s mobile phone. The taxi-hailing software is divided into three, six, nine, etc. The more expensive the mobile phone, the more expensive the taxi. Big data judges your spending power based on what brand of mobile phone you use, and even tailors the charging standard for you. Is this the Internet of thousands of people? Figure / People’s website official micro Do you think how much battery power the phone has is privacy? Most people may say no. But in fact, the battery level reflects a very important information-if the battery level of the mobile phone is always full or increasing, it means that the mobile phone is in a charging state, and the user is most likely to be indoors. In addition, the sensor gyroscope used in the mobile phone to measure our pace and determine the motion posture can determine the user’s motion status. Whether you are walking, running, driving or taking the subway, you can’t escape its “legend”. “If the battery of a mobile phone is always full, it can be judged that it is plugged into the charging line, and it can be judged from the speed that it is driving, and this state can last for ten hours a day, then this can be judged. People’s profession is a taxi driver.” “For another example, do you think it is possible for the mobile phone system to know your work location and home address? If you want to know where you work, the system only needs to sample you once every morning at 10 o’clock. If it finds that there are 20 days in a month If you are all in the same place, then you can basically determine your working place. Similarly, if you put the sampling time at 12 o’clock in the evening, you can draw a conclusion about where your home is.” Liu Peng passed two simple The example illustrates how big data knows personal information well. In a sense, today’s mobile devices have become an “organ” of you. Apart from sleeping, you can hardly imagine any scene where you can do without a mobile phone. Since it is an organ, it may leak a lot of your privacy. After all, there are a lot of sensors on the mobile phone, which can understand a lot of personal information. Some information can be obtained in ways that non-professionals can’t even imagine. “What I want people to know is that everything they do online is monitored, tracked, and evaluated. Every behavior you make is carefully monitored and recorded, down to where you stop. A picture may be recorded for as long as it has been viewed.” Jeff Seibert, a former Twitter executive, said in the documentary “Surveillance Capitalism: Smart Trap”. The algorithm may know yourself better than you. It knows your emotions, whether you are introverted or extroverted, and holds all your personal information. What can be more frightening than this? 2. Fall into the “algorithm trap” It is true that some people say that the things recommended by the algorithm are more in line with their own hearts and save the time of selection; but some people say that we have just fallen into an “algorithm trap”. Let’s take a look first, how the algorithm understands our preferences more intelligently and makes corresponding recommendations. “You can judge whether you are a price-sensitive person through data. Take e-commerce as an example. Many sellers will open more than a dozen stores on one platform. The same product sells at different prices in each store. Then, according to your spending power, The algorithm will recommend you the price you can afford. Price discrimination is impossible to eliminate, which violates the nature of business, but it is easier to implement it under the conditions of the Internet.” Liu Peng told the explorer ” Tanker”. After technology companies collect user data, they will make a model to predict user behavior. The more data the model will become more accurate. Every time everyone clicks on the screen, whether it’s likes, disinterested, likes, etc., these data will become a database for shaping the model tailored for you. After all, in the market competition, once more user data is obtained and a more accurate model is made, the company has the opportunity to occupy a larger market space. Photo / A station “Surveillance Capitalism: Smart Trap” documentary “The recommendation logic of the Internet mainly depends on what your objective function is. For example, content recommendation, whether to make the click rate higher or the browsing time longer, or to make the user more satisfied, the objective function of the recommendation logic will be different. It’s different. After constant positive feedback, the accuracy of the algorithm will become higher and higher, and the recommendations will become more and more accurate.” Liu Peng introduced to the explorer “Tanker”. The emergence of the Internet has made many people feel convenient, but technology is a double-edged sword. The more the system knows about you, the more accurate the things recommended by big data will be, which will allow you to spend more time with them. The application makes you more and more addicted to mobile phones. In this regard, the current popular social software also uses recommendation algorithms to attract the attention of users, and ultimately make users fall into a state of “more and more brushes”. For example, short-term signals such as likes and comments on social platforms can give people psychological satisfaction. This feeling of satisfaction is a kind of “reward” for people. After receiving the reward, people will look forward to the next experience. There is greater satisfaction. This is one of the important reasons why people are addicted to social networks. “There has been a substantial increase in depression and anxiety among the American youth group. The number of teenagers who self-harm or even commit suicide and admitted to the hospital has risen sharply. The proportion of girls aged 10 to 14 who have self-harm behavior has even increased by 151%. This growth pattern points to social media. .” said Dr. Jonathan Haidt, a social psychologist at New York University’s Stern School of Business. This is another aspect of society as a whole falling into the “algorithm trap”. Children born after 1996 have started to use social media from the junior high school stage, and they are also called “Internet natives” after 00. These young people are more familiar with Internet life and are more prone to anxiety, vulnerability or depression due to the Internet environment. mood. Previous surveys have shown that many young people have “mobile phone anxiety”. Once the mobile phone is not at hand, they will fall into anxiety and feel at a loss. To a certain extent, one of the reasons for the emergence of this new type of symptom is that the algorithm attracts people to immerse themselves in mobile applications, which makes people feel “happy to play with mobile phones”. Mobile phones make life more convenient, and algorithms make life easier. People are more immersed in it, and ultimately people cannot do without the virtual world on their mobile phones. 3. Correctly understand the recommendation algorithm Objectively speaking, any technological innovation cannot escape the coexistence of advantages and disadvantages, and the same is true for recommendation algorithms. “First of all, we cannot demonize it. If you cannot understand your preferences through data today, the information flow you receive will be very chaotic, and the experience will be worse. People are worried about privacy being used, but in fact, the purpose of using data is Just to provide you with personalized content and advertisements, users actually have no loss, and the platform will not exchange or sell your data. From the perspective of economic benefits, it is not worth it for them to do these things. This is put into production. The output ratio is unbalanced.” Liu Peng explained to the prospector “Tanker” from the perspective of the enterprise. In fact, our country has certain standards for the use of data, and some boundaries are not touchable, such as the user’s phone number, E-mail, and ID number. For many people, the risk of data leakage does not come from these large platforms, but may occur on some small platforms, or individuals in the company, for example, an employee who can see the data steals your courier address from the database. Pour it out and sell it. “It is impossible to completely solve the problem of data protection through regulations. More technical means are required. For example, the core fields are encrypted to ensure that internal engineers can’t see the data.” Liu Peng told the explorer “Tanker” wanted to be true To solve the problem of data protection, we still need to rely on technical means. At present, on the issue of data security protection, companies such as BAT have teams doing research and development of related encryption technologies. In addition, there is another question that has plagued everyone for a long time-will the application really listen to us? Is this a means of recommending algorithms? “This is possible in principle, but the practicality is very poor. If the software wants to listen all the time, you need to turn on the recording function inside the App. In this case, the power consumption is very high, and your phone will soon No battery, I personally think it is impossible, if it is text and input method, it is possible.” Liu Peng answered this question for us from a professional perspective. If the user is still worried about this issue, perhaps consider turning off the microphone permission of some software in the phone settings. In addition, many users still have a question about the recommendation algorithm of the e-commerce platform-why do we always recommend the things we have bought? Doesn’t the algorithm know that we have already bought this product? Is this a performance that it is not smart enough? “In fact, this is because many users have a high repeat purchase rate. From a technical point of view, it is easy for the platform to stop the algorithm from recommending what users have purchased, but from a profit point of view, repeated recommendations are very effective.” Liu Peng said to the detective “Tanker”. In essence, the data and algorithms themselves are to enable the platform to more accurately push to target users, so that users have a better experience, but in the process of technological development, various situations will be encountered. Different companies Different recommendation logics will also be formulated according to different needs. Perhaps an algorithm will lead to the polarization of society, and even make you manipulated, unable to escape from the model created by the algorithm for you, and fall into the “information cocoon.” But to a certain extent, algorithms are also facilitating our lives. What we really need is that the technology industry develops better technical means to circumvent existing problems. At the same time, people who formulate and run algorithms have higher ethical requirements for their behavior, and policies and regulations have stricter regulations for people’s data security. With protection, users can have a broader vision and the ability to think independently. It can be seen that this is still a problem for the whole society and requires everyone to work together to explore and improve. Just as the problem of big data has been raised by people, it has the possibility of being solved and optimized. In the final analysis, what we want to see in the future is that recommendation algorithms are not only smarter, but also more virtuous.