How does the Google search engine actually work?
While this is a subject entire books have been written about, there's a good chance you're in the market for something a little more concise. So let's say it's getting close to dinner and you want a recipe for lasagna. Google has been continuously mapping the web hundreds of billions of pages to create something called an index. Think of it as the giant library Google looks through whenever you do a search for lasagna or anything else now the word lasagna shows up a lot on the web pages about the history of lasagna articles by scientists but if you're hungry randomly clicking through millions of links is no fun. This is where Google's ranking algorithms come into play. First they try to understand what you're looking for so they can be helpful even if you don't know exactly the right words to use or if your spelling is a little off. They search through millions of possible matches in the index and automatically assemble a page that tries to put the most relevant information up top for you to choose from.Now we have some results but how did the algorithms actually decide what made it onto the first page? There are hundreds of factors that go into ranking search results. pages containing the words you search for, are more likely to end up at the top, no surprise there but the location of those words like in the page's title or in an image's caption. Those are factors too. There's a lot more to a ranking than just words. Back when google got started they looked at how pages linked to each other to better understand what pages were about and how important and trustworthy they seemed. Linking is still an important factor, another factor is location where a search happens because if you happen to be in italy you might be looking for information about their annual lasagna festival but if you're in omaha nebraska you it becomes irrelevant.
Steps involved in Google Search Algorithm
Google search works through a complex system called the Google Search Algorithm, which follows these primary steps:
Crawling: Google uses automated programs called crawlers or spiders to discover web pages. These bots follow links from page to page, gathering information about each webpage they visit.
Indexing: After crawling, Google organizes and stores the information from web pages in a massive database called the index. This index contains a vast collection of web pages and their content.
Ranking: When a user enters a query into the search bar, Google's algorithm sifts through the index to find relevant pages. It then ranks these pages based on hundreds of factors such as relevance, quality of content, freshness, authority, user experience, and more.
Displaying Results: The search engine presents the most relevant and useful results to the user in the Search Engine Results Pages (SERPs). The results are displayed based on their ranking, aiming to provide the most accurate and helpful information for the user's query.
Google's algorithm continuously evolves and updates. It refines its processes through machine learning, AI, and user feedback, aiming to improve the quality and relevance of search results.
Google also provides various features and enhancements to its search results, including
Snippets (short excerpts of information from a webpage)
Knowledge Graph (displaying information in panels for quick reference)
Rich snippets (enhanced search results with additional details like star ratings or prices),
In order to enhance the user experience and provide more direct answers to queries when possible. When a web page was uploaded is an important factor too pages published more recently often have more accurate information especially in the case of a rapidly developing news story. Of course not every site on the web is trying to be helpful just like with robocalls on your phone or spam in your email. There are a lot of sites that only exist as a scam and every day scammers upload millions more of them. Google spends a lot of time trying to stay one step ahead of tricks. Making sure their algorithms can recognize scam sites and flag them before they make it to your search results. So google reviews billions of times a day whenever someone searches for lasagna or resume writing tips or how to swaddle a baby or anything else. Google software locates all the potentially relevant results in the web, removes all the spam and ranks them based on hundreds of factors like keywords, links location and freshness.
The web is always changing and people are always searching for new things, in fact one in every seven searches is for something that's never been typed into the search box before. So google is always working and updates to search thousands every year which brings up a big question, how do we decide whether a change is making search more helpful? Well, one of the ways we evaluate potential updates to search is by asking people like you. Every day thousands of search quality raders look at samples of search results side by side then give feedback about the relevance and reliability of the information. To make sure those evaluations are consistent the readers follow a list of search quality evaluator guidelines and think of them as publicly available guides to what makes a good result. One last thing to remember google uses responses from readers to evaluate changes but they don't directly impact how search results are ranked so there you have it every time you click search our algorithms are analyzing the meaning of the words in your search matching them to the content on the web understanding what content is most likely to be helpful and reliable and then automatically putting it all together in a neatly organized page designed to get you the info you need.
How BERT helps google’s search engine
If a pancake recipe told you to mix the butter with the banana you probably wouldn't think to use the banana as a mixing spoon but what's obvious to humans things like context tone and intention are actually very difficult for computers to pick up on. At its core a google search is about understanding language in order to return the right information. Google doesn't just need to know the definition of the words, it needs to know what they all mean when structured together in a specific order. This includes the smaller words like four and two and when you think about how many different meanings a single word can have you start to see how writing a computer program that takes all into account is pretty tough. More people talk to google the way they think and speak and more and more google is getting better at understanding what they mean. One of the biggest leaps forward in the history of search came about with the introduction of bidirectional encoder representations from transformers or as we like to call it BERT. BERT is a machine learning model architecture that helps google process language and understand the context in which it appears in search engines. There's a big difference between knowing words and understanding meaning. The model learns context by applying the same fill in the blank principles it takes to complete a mad libs so we take a phrase we hide about twenty per cent of the input words and then we make the computer guess the words that are missing over time. The model begins to understand different words have different meanings depending on what's around them and the order in which they appear in that text really matters so when you search something complex like fly fishing bait to use for trout in september montana. Search engines knows all the little words are important and because it now takes them all into account it can tell you the perfect bait for that time of year.
Bình luận