Since we got introduced to RankBrain (Google's machine learning artificial intelligence system) two years back, we've seen the fast-paced updates on Google's Algorithm and how much the search landscape has changed. With user experience as its main goal, Google continues to evolve web search in a myriad of ways including artificial intelligence. From refining better web experience for desktop users, the focus was diverted to web mobile. Later on, search engine algorithm came up with methods to analyse each differently based on the premise that there are hidden search intent on top of the more obvious that it intelligently attempts to uncover through machine learning.
Back in October 2015, RankBrain was already handling a large portion of millions of queries that go to Google each passing second. In a nutshell, we will describe how it works in the best way we can. Machine learning is an artificial intelligence capability that gives computers the ability to learn without being explicitly programmed. RankBrain uses a machine learning technology to analyse and process search queries. If it sees an unfamiliar word or phrase, it is capable of guessing by relating it to other words or phrases with related meaning and return highly relevant content even if the results do not contain the search keyword or search phrase. This means that content writers no longer have to force a keyword repetitively and can now focus on writing naturally.
RankBrain also analyses the search intent that goes beyond the more apparent initial content by digging deeper - such as a possible problem that a visitor may be trying to solve during the time of search (so time here is a factor). Others may include whether the visitor is browsing through web desktop or mobile, gender, location, relation to previous related searches, current events, etc. There may be lots more that can be added to the list as the machine learning capability evolves. For example, if a male searches for a bar on his desktop at 5 PM on a Friday may be looking for a place where he and his drinking buddies can get a booze to cap the week. Searching for a bar at 7 PM on his mobile means he could be out of the office already and look for a place nearby. His responses and impressions to results from past searches could also be considered to generate more relevant search results based on preferences. So here, we can infer that there is deeper thought on what a visitor's underlying intent could be and what his secondary intents are as well.
So we've seen so much of machine learning and artificial intelligence with Google. In its recent AI event in San Francisco California, Microsoft presented its vision of computing and differentiation strategy with the help of Artificial Intelligence. The talk touches largely on creating AI-supported software that is highly accessible so people can use it to improve experiences every day, the ability for users to combine work and personal functionality and becoming an ethical AI company. Apparently, some of its products are already AI-supported - from Windows to Office 365, but Bing tends to be a major highlight. Artificial intelligence announcements about Bing centred on "intelligent search" - intelligent answers, intelligent image search and conversational search.
Intelligent search is comparable to next-generation featured snippets wherein Bing summarises and compares multiple sources of information instead of presenting only one answer. Competing answers will be presented, and multiple answers to a question will be shown in a carousel of intelligent answers.
Intelligent image search is similar to Pinterest's visual search and object recognition. This feature is currently focused on fashion and home furniture. By clicking the magnifying glass icon on the top right of any image, users can search within an image to find related images or products. Another image search feature is its ability to identify buildings and landmarks in user photos or image search. We anticipate that Bing will soon have the capability to identify images in the real world as Google Lens does with objects and places and Amazon for products.
Another interesting, intelligent feature of Bing is conversational search. This is the ability of Bing to help users refine vague search queries to get the best answers the first time. Users will initially experience this for searches on health, tech, and sports queries and will be expanded over time. Since these features are built on large-scale machine learning, Bring promised that the experience would enhance over time.
0 comments:
Post a Comment