Introduction to Knowledge Graph

Have you ever thought how the search engine giant GOOGLE manages to show us such a precise and vast list of results from around the world web for whatever keyword we type-in to search?



I thought when a few days back my Team Leader showed me a video of Google's I/O 2016 which made me think about what great and amazing work they are doing for us. But the question that aroused in my mind was HOW?

So I went on to google it how Google is doing what it's doing. I know the answer is not limited to just one line or one strategy or one algorithm but there would be thousands of queries and search algorithms working behind for the results we get only in seconds. To be specific google processes over 40,000 search queries per second on an average which serves to over 3.5 billion searches per day. Isn't it amazing?

But the question is where all these data come from? Google has a huge knowledge base where it stores all the information gathered about its users and their general trends like what they search, look, do, and use. Other than extracting data from Wikipedia, Freebase and CIA World Fact-book, there are numerous techniques working for such large but close data collection of the user's general interests and trends such as searches we made, web crawling, web analytics, ad servings, web clicks, google apps that we use like Gmail, youtube, twitter, feed burner, google books, google news and so on. Google has spotted various data collections to support its knowledge base. It acquired the public online collection Freebase to enhance its search in 2010. And after that, there comes Google's new idea of KnowledgeGraph that replaced the Freebase to make searching a totally different experience for its users and I guess this is what we are having today.


In 2012 Google and Freebase team decided to shut down Freebase and adapted a new technique of searching to make it a real-world search engine. The conventional searching includes matching the keywords to the queries and indexing into the data pool to return the most relevant results. But the Knowledge graph is somewhat different. It is a more semantic, fast, and enhanced way of searching. Unlike conventional databases, the Knowledge Graph is not just about storing the object but instead storing and understanding the facts and relationships between those real-world entities to refine the results returned to the user query.


The main idea behind this new initiative was to make an intelligent and smart system you can talk to and it will answer everything you ask it. And in order to complete this dream, Google is making continuous efforts and has almost done by announcing its new product Google Home which is also based on the idea of understanding and answering your real-world queries and act accordingly to help you in your daily busy daily life. The knowledge graph is about the actual knowledge and it is knowledge with some other knowledge. The knowledge graph understands the entity and provides the relevant data with some other information of interests. We're all familiar with the “People also search for” feature of google search. While writing this blog I searched for “Google” and it showed me People also search for Facebook, Twitter, LinkedIn, Google+, and Instagram. Similarly, I searched for Larry page and I also got to know about other CEO and founders like Sergey Brin, Sundar Pichai, Lucinda Southworth, Mark Zukerberg and Eric Schmidt. The knowledge graph provides search results for all your queries regarding places, peoples, buildings, books, celebrities, sports, whatever google knows in a more intelligent way.


The knowledge graph is smart enough to return the most relevant result to your query whether you're searching for Taj the beautiful monument, Taj the famous hotel, or Taj the casino. I've previously said the knowledge graph is knowledge with some other knowledge. That is true. It is actually a graph of related knowledge and information about the entities it contains, to make it really quick and easy to get the most and best out of a search as it summarizes the relevant content around the topic. For example, if I search for Sunder Pichai on google, I also get to see on the side panel of the search page a graph of his personal info like when he was born, what is his nationality, education qualification, his parents and spouse, etc. If I google about one famous author or singer I'll also get to know about other authors or singers which actually add to my GK ;-).


And Google is not the only one using this technique but the other social giants and competitors like Microsoft, Yahoo are also working on it, but Google's knowledge graph is not only dominating the internet and market but it's even stabbing or say killing the Wikipedia. Google is really doing a remarkable job and we hope to see some more amazing and useful inventions from Google in the near future that will become a trend for all the tech giants. Behind a simple page of search, a really complex system carefully crafted, structured, and maintained is working to provide us with the best results possible. And while you were busy reading this post, in the meantime there have been performed another billion searches. So enjoy googling with Google's Knowledge Graph. ;-