Let me tell you another famous technology.
Of course, the most famous and patented technology is Google's PR technology.
You may have heard that Google's revolutionary invention is its web page ranking algorithm called "page ranking", which completely solves the ranking problem of search results. In fact, it is not the first time that Google has tried to rank many websites on the Internet. Yahoo! The company first let users search for information on the Internet through directory classification, but due to the limitation of computer capacity and speed at that time, Yahoo! Other contemporary search engines have a common problem: they contain too few web pages and can only index the actual words related to common content in the web pages. It was difficult for users to find very relevant information at that time. I remember that before 1999, I had to change several search engines to find a paper. Later, DEC developed AltaVista search engine, which only used an ALPHA server, but contained more web pages than the previous engine, and indexed every word in it. Although AltaVista allows users to search a large number of results, most of the results have nothing to do with the query, and sometimes it takes several pages to find the web page they want to see. So the original AltaVista solved the coverage problem to some extent, but it could not sort the results well.
What about Google's "page ranking"? In fact, it is a simple democratic vote. For example, when we were looking for Dr. Kai-Fu Lee, a hundred people raised their hands and said they were Kai-Fu Lee. So who is real? There may be several, but even so, who are you really looking for? :-) If everyone says that Google is true, then it is true.
On the Internet, if a webpage is linked by many other webpages, indicating that it is generally recognized and trusted, then its ranking is high. This is the core idea of page ranking. Of course, Google's page ranking algorithm is actually much more complicated. For example, the links of different web pages are treated differently, and the links with high ranking of their own web pages are more reliable, so these links are given greater weight. Page Rank considered this factor, and now the problem is coming again. In the process of calculating the ranking of search results, you need to rank your own web pages. Isn't this a chicken or an egg problem?
Larry, two founders of Google? 6? 1 Larry Page and Sergey? 6? Sergey brin (1) turned this problem into a two-dimensional matrix multiplication problem, and solved this problem by iterative method. They first assume that the rankings of all web pages are the same. According to this initial value, they calculate the first iteration ranking of each web page, and then calculate the second ranking according to the first iteration ranking. The two of them proved theoretically that no matter how the initial value is selected, this algorithm ensures that the estimated value of web page ranking can converge to their true value. It is worth mentioning that this algorithm has no human intervention at all.
Theoretical problems have been solved, but practical problems have been encountered. Because of the huge number of web pages on the Internet, the two-dimensional matrix mentioned above theoretically has multiple elements with the square of the number of web pages. If we assume that there are one billion web pages, then this matrix has one billion elements. Multiplication of such a large matrix requires a very large amount of calculation. Larry and Sergey use the skills of sparse matrix calculation to greatly simplify the calculation and realize this web page ranking algorithm. Nowadays, Google engineers transplant this algorithm to parallel computers, which further shortens the calculation time and makes the cycle of web page update much shorter than before.
After I came to Google, Larry talked with several of our new employees about how he and Sergey came up with the webpage ranking algorithm. He said: "At that time, we felt that the whole Internet was like a big picture, each website was like a node, and the links of each webpage were like an arc. I think the internet can be described by a graph or matrix, and I may use this discovery to do a doctoral thesis. " So he and Sergey invented Page Rank's algorithm.
The genius of webpage ranking is that it treats the whole Internet as a whole. It unconsciously conforms to the viewpoint of system theory. In contrast, most of the previous information retrieval regarded each web page as an independent individual, and many people only paid attention to the correlation between the content of the web page and the query sentence at first, ignoring the relationship between the web pages.
Today, Google search engine is much more complex and perfect than the original one. But page ranking is still very important in all Google algorithms. In academic circles, this algorithm is recognized as one of the greatest contributions in literature retrieval, and has been introduced into information retrieval courses by many universities.