Because of the great success of Artificial Intelligence (A.I.) in almost all domain of applications, big giants like Google, Facebook, Twitter, Intel, Apple are in a rush to grab Artificial Intelligence Startup. They are investing billions on Research and Development in Artificial Intelligence applications and products. This shows great importance of A.I. People may be curious to know about future of Google Search – Artificial Intelligence. The article briefly explain how Google is using Artificial Intelligence in its applications.
RankBrain – Google’s Artificial Intelligence
Every day there around 3.5 billion searches on Google, out of which small percentage of queries have never been fired. To handle these unknown searches, Google has developed, RankBrain, an artificial intelligence based system. In addition to finding the most appropriate sites accurately to match the query, it also learns from its mistakes to make future search results better.
In a tests, Google engineers correctly ranked 70 per cent of sites while RankBrain attained a score of 80 per cent. Google handles the ever increasing size of search traffic using Artificial Intelligence. RankBrain also contributes to the ranking of pages.
This Artificial Intelligence approach, called deep learning, is a deep neural networks. It approximate the web of neurons in the human brain. By analysing vast amounts of digital data, these neural nets can learn all types of useful tasks, like recognising commands spoken into a smartphone, identifying photos, and, as it turns out, responding to Internet search queries.
If you feed enough photos of a tiger into a neural net, it can learn to identify a tiger. If you show it enough computer virus code, it can learn to recognise a virus. In some cases, they can learn a task so smartly that they surpass humans. They can do it faster, better and at much large scale.
2. Google Translate – A Translation Service
Google Translate is a language translator based on Artificial Intelligence algorithm. The company is constantly charging its translation application using artificial intelligence (AI). It is using Neural Machine Translation into Google Translate, which has radically improved results. AI team of the company calls it the Google Neural Machine Translation system (GNMT).
GNMT takes in the entire sentence as a whole, instead of each word or phrase as a standalone unit. The benefit of the approach is that it requires fewer engineering design choices than previous phrase based translation.
This new method uses a total of 16 processors to first transform words into a vector. It represents how related one word is to every other word in the vast dictionary of training materials. For example, “cat” is more closely related to “dog” than “bike,” and the name “Narendra Modi” is more closely related to “Soniya Gandhi” than the name for the country “America”. The method uses vectors from the input language to come up with a list of possible translations that are ranked based on their probability of occurrence.
When the application was first introduced, its accuracy was matching the existing translation systems. However, over time, GNMT has proved capable of both producing superior results and working at the speed required by Google’s consumer apps and services.
Thus Google is essentially powering their application with machine learning algorithms and day by day getting better and better results. It is said that, Artificial Intelligence is the future of Google Search, and if it’s the future of Google Search, it’s will also be future of so many more. The company has achieved great success in RankBrain and many other AI applications. It has invested billions for research in Artificial Intelligence because of its amazing outcome. Therefore, the statement “Future of Google Search – Artificial Intelligence” is gradually becoming true.