New careers for large data and artificial intelligence: algorithmic Division

You should remember that the "Matrix" Leo to avoid bullets in the highly difficult action, the film, full of large data and artificial intelligence color. The algorithms are now widely used because they are marked with database, detected, choreographed, and even attacked by the enemy.

Large data is the ability to quickly obtain valuable information from a variety of types of data. It is important to understand this and it is this that has the potential to move to many businesses.

From Amazon's price recommendation, Nefilix's successful plan for Solitaire, and David Rothschild, an economist at Microsoft's New York Institute, for accurate predictions of Obama's Oscar-winning NBA data, has amply demonstrated that this is not a gimmick.

Large data has four distinct advantages: first, the volume of data is huge. Jump from TB level to PB level; second, a wide variety of data types. The aforementioned blog, video, pictures, geographic information and so on. Third, the value density is low. In video, for example, a continuous uninterrupted monitoring process may be useful for only two seconds. Four, the processing speed is fast.

Artificial intelligence is a very complex direction, is also a interdisciplinary computer science, involving other aspects of knowledge, the main content is computer learning human natural language processing. We don't have to daydream about sci-fi like a smart Terminator, that's what scientists are going to study. In real life, we've seen more and more large numbers of products with artificial intelligence, such as Apple's Siri, which contains speech recognition (Speech recognition) and is part of natural language processing (Natural Language 處理). In the actual application, played a lot of role, including the smart phone service deployment, can solve the user's daily needs.

So is this a combination of big data and artificial intelligence? Let's look at the simple principle:

In the book Big Data: Revolutions in our lives, jobs, and Reflections, the author Meyer, the current computer system is based on rules that explicitly require them to follow when writing programs. Therefore, when a result is occasionally unavoidable error, we can go back to Recode. No matter how complicated the computer code is, any code is the basis of the operation that can be pursued and understood.

But the tracking of big data has become much harder. First, the foundation of algorithmic predictions may be so complicated that it is difficult for ordinary people to understand. Google translation uses billions of of pages of translation to judge a Word's translation. This statistical operation based on massive data makes it almost impossible to trace the specific factors of the algorithm. At the same time, because the size of large data, the size of its operation is beyond our imagination. Google's identification of several search keywords and flu links is the result of testing 450 million mathematical models.

If you want to persuade customers to use this technology, you need algorithms to help adjust, the person in this profession, what needs?

First, these professionals are experts in computer science, mathematics and statistics. In their daily work, they examine the analysis and prediction of large data. They will evaluate the data source, analyze the forecast, make the underlying algorithm model, when people need to detect the principle of the algorithm, they will pull out the algorithm results, statistical methods and databases. To put it simply, the Algorithmic Division is responsible for a function of filtering data.

Computer development So far, the accumulation of huge databases need someone to organize their organization, give targeted use. Here the algorithm division is divided into external and internal algorithm master. External algorithms can examine the accuracy or reasonableness of large data forecasts as a neutral auditor, when the government needs them, such as by issuing orders or enacting regulations. Algorithmic Division can provide services for large data companies and give professional audit services.

The interior is the master of the internal monitoring of large data activities. They not only pay attention to the interests of enterprises, but also to the enterprise's large data analysis of the impact of the interests of the people. They oversee large data operations, and when anyone feels they are being hurt by the agency's big data, the in-house algorithmic division is the first person to contact. They check the completeness and accuracy of the analysis before the data is released. To accomplish the first two tasks, the algorithmic division must enjoy a certain degree of freedom and neutrality in the enterprise in which they are serving. In short, the internal algorithmic division is the enterprise to maintain public trust and health of the occupation.

The most direct reason for this professional demand for algorithmic division is that the field of large data has not yet established a new normative standard to restrain the enterprise. Algorithmic division through the design of a system for the social analysis of personal data and other security concerns to set up security, for this open the black box occupation, someone interested?