American scholars question "Big data" theory

Editor: The 2012 end of the "Big Data Age: Life, work and thinking of the great changes," with a new concept of predictability, caused by the global hot debate, still living in the rankings of major books, author Victor Maire-Schoenberg also known as the "Big Data Age Prophet." Public opinion, "large data" subversion of the thousands of years of human thinking practices, human cognition and communication with the world put forward a new challenge. At the same time, because of "big data" theory triggered by the debate and questioning also a stream of ...

Two of America's most prestigious academic journals have also focused on big Data topics: "The rise of Big Data", a cover article in the 5/6 Journal of Foreign Affairs, argues that big data will change the way people think and look at the world. The foreign Policy magazine on the Web site published the lead researcher of Microsoft Research, the MIT Civic Media Center, visiting Professor Kate Claufford's article "On large data Rethinking," from five aspects of large data theory questioned.

Big Data brings change

"The rise of Big Data" is the author of Kenneth Couqueil and Victor Maire-Schoenberg, author of the recent Hot discussion monograph "Big Data Age: Great changes in life, work and thinking". In the article, Couqueil and Schoenberg affirm the great change ability of large data to society, the pro data not only will change people's life and work, but also will change the way that human understand and think the world.

Two people believe that, as the technological environment changes, on the one hand, the world "data explosion" phenomenon, on the other hand, the ability of human processing data is greatly enhanced. As a result, there are three changes in the way people treat data: First, the data that people processed changed from sample data to full data; Secondly, because it is full sample data, people have to accept the mixed data, and give up the pursuit of accuracy. Third, the human through to the large data processing, abandons to the causal relation the desire, turns and focus on the interconnectedness. All this represents the attitude of humanity to say goodbye, always trying to understand the underlying causes behind the workings of the world, and moving towards merely clarifying the link between phenomena and using that information to solve problems.

The rise of large data lists examples of applications of large data in the fields of medicine and consumer goods. But the authors also argue that the impact of big data is by no means limited to business, and that it will profoundly change the way governments operate and the nature of politics. "Those who can effectively use large numbers will have a huge advantage over others in promoting economic growth, providing public services or fighting wars," they wrote. However, the authors acknowledge that the more successful cases of large data applications in public services appear at the city level because it is easier to obtain data and use information at this level.

Scholars put forward five questions

Crawford's article argues that big data is the current buzzword, but it is doubtful whether people can rely on massive amounts of data to reveal the laws of human behavior. She questioned the big data theory from five aspects.

First, there are biases and blind spots in large data. Advocates of big data say that "with enough data, numbers can speak for themselves". But Crawford believes that figures cannot speak for themselves. Regardless of their size, datasets are ultimately the product of human design, and the tools of large data do not make people free from distortions, estrangement and false stereotypes. Recognizing these factors is particularly important when large numbers attempt to reflect the social world in which humans live. Biases and blind spots exist in large data, and conclusions from large data are no more objective than man-made opinions.

Second, large data can make cities more intelligent and efficient to some extent, but how the effect depends on the knowledge of the data and its limitations by municipal officials. "Big data will make our cities smarter and more efficient," says Crawford, a big-data advocate, to some extent. On the other hand, data is not equal in the process of generating or gathering, and there is a "signal problem" in large datasets, that is, some people and communities are neglected or insufficiently represented. Therefore, in order to use large data, municipal officials must have a good understanding of the data and its limitations.

Third, large data can lead to discrimination on the basis of groups. Large data advocates argue that "big data does not discriminate against different social groups" on the grounds that the analysis of raw data is conducted on a large scale, thus avoiding discrimination on the basis of groups. But Crawford believes that is not the case. Since large data can make assertions about the different patterns of behaviour of groups, and the main purpose of their use is to classify different individuals into different groups, large data will not only avoid group discrimination, but may also aggravate this trend.

The disclosure of privacy is an important problem in the application of large data. Crawford argues that big data advocates say "Big data is anonymous, so it won't violate our privacy" is a mistake. Although many large data providers try to eliminate individual identities in the data, the risk of identity being reconfirmed is still high. Given the amount of information that can be inferred from the use of a large number of public data sets, revealing personal privacy becomes a "growing concern".

The big data provides a new way for scientific research, but it is not asserted that "big data is the future of science". The research method of large data can only count the frequency and relevance of something, but it does not produce causation, Crawford said. Combining large data strategies with small data research may be a better way of scientific research.

Couqueil and Schoenberg also recognize some of the inherent flaws in the big Data theory. At the end of the article "The Rise of Big data," the authors say that large data is a resource and a tool for informing, not interpreting; it is intended to promote understanding, but the key to misunderstanding is the degree to which people have mastered it. They believe that people must accept big data in a manner that appreciates not only its power but also its limitations.

(Ludo Bao)

(Responsible editor: The good of the Legacy)