Kelly observation: Third Generation search

What is the third-generation search?

What is the difference between the third-generation search engine and the first-and second-generation search engine? This is the first question I asked him before he opened his notebook.

The climax held that the first generation of search engines used reverse indexing technology, mainly using search keywords. However, the first generation of search engines is mostly weak due to the popularity of websites. Therefore, the second generation of search engines use keyword-based website link analysis to achieve search. The third-generation search engine searches for sentences and phrases. This is almost a realm of exhaustive search needs. That is to say, it is impossible to generate the fourth generation of search technology.

What are the third-generation search technology trends?

According to the climax, including Google, Microsoft, and many professional search companies, the development of third-generation search engine-related technologies and products is in progress, and there are no successful products yet, let alone products.

A company named senopy is using natural language to develop search engines, but the speed is slow to the point where users are unbearable (generally, users' waiting time is about seconds ). There is also a content-based search engine developed by trovix.com that needs to be implemented offline. A typical application case is the resume of the job seeker corresponding to the job requirements. The climax found that the mature gene Sorting Technology and variable length, variable interval technology into the search engine, can be based on the content of the search intelligence, and the speed is improved by thousands of times.

Genetics and Chinese

Genes are sorted by four nucleic acids, proteins, and 20 amino acids. There is no interval between them, and there is no interval between Chinese words and words (there is an interval between English words and words ). To search based on phrases and sentence content, it is necessary to accurately identify the variable length and variable interval of sentences. The latter is more difficult to identify. For example, Chinese and clothing are two different topics. The variable interval in English can be achieved by the existing Word Segmentation: I like movies and I like action movies very much. I like movies and I like kung fu movies very much. These two sentences fully show word segmentation with Variable Length and variable interval.

Why makes third-generation search faster

Why is the climax of third-generation search times faster than traditional search? This is based onAlgorithm:

For example, 10! = 36288002> 2*(10/2 )! = 240, the latter is obviously more than 1000 times smaller than the former.

Isn't it easy to imitate? No, the climax said that if variable interval recognition is not achieved, it would be impossible to complete Quick Content Search Based on phrases and sentences.

Traditional search engine algorithms use keywords as vector coordinates, while third-generation search uses phrases and sentences as vector coordinates.

Partners needed for the climax

After reading the climax of the search demonstration, I felt that the previous introduction was not just about talking on paper. His search was not only fast, but also based entirely on phrases and sentences.

If the third generation of search engines are completely commercialized, all searches will turn into computer-to-human communication and conversations. All the movies adapted from science fiction have become reality, and the whole society

Everything changes. It can be said that the climax of research into the computer into the human brain has produced a secondary business-the third generation of search support.

Conclusion

After 21 years as a reporter, I wrote a search engine for the first time. I knew that some technical problems were not clearly understood, but I knew that many readers were better at understanding it.