AI is proving itself to be a powerful technology as each day progresses by.
At its Search On event, Google apparently announced a lot of alterations to its Google search service. These changes will be put into effect within a few weeks or months.
The main reason for these changes is to provide better search results using AI and machine learning techniques to the users. Here is the biggest advancement. Google has unveiled a new spelling checker tool that deciphers even the most poorly spelled words.
1 out of 10 queries are misspelled everyday
According to Google’s head of search, Mr. Prabhakar Raghavan, around 15 % of the queries every day are those that Google cannot identify at all. This means that the company has to work further to provide better results.
The main reason for that is misspelled queries.
The VP Engineering at Google, Cathy Edwards says that 1 out of every 10 queries on Google are spelled wrong. Google tries to help us out with its did you mean feature that indicates the correct spellings. This feature will get a brand new upgrade by the end of the month. This update will feature a new spelling algorithm fueled by neural net along with 680 million parameters. Here, the algorithm runs within 3 milliseconds after each search. Google promises that it will provide a better search recommendations for misspelled words and queries.
The upgrades are not yet over. Here is another one.
Google can now index passages separately from webpages
Instead of the whole webpage, Google can now mark individual passages from webpages.
For instance, let’s say you type “how can I determine if my house windows are UV glass” into the search box.
The latest algorithm can find a single paragraph from the DIY forum instead of the whole webpage. Finding answers will become more simpler and easier. Edwards states that, when the feature is put to action, it will improve 7 % of the queries across all languages.
Additionally, to provide accurate results, the company is using AI to divide the broader searches into sub topics.
Other latest technologies like computer vision and speech recognition are being used too. These are used to automatically divide and tag videos into parts and segments. Let us take cooking videos for instance. They will be automatically parsed and divided into chapters.