How the search engine uses MUM and BERT for safer search results.

2 min read
How google interprets search query

Do you ever wonder how Google interprets your search queries? There’s a lot that goes into providing useful search results, and one of the most critical aspects is language interpretation. Search systems understand human language better than ever before, thanks to advances in AI and machine learning. And Google recently looked at the scenes at how this translates into relevant results for you.

Google’s VP of Search shares new information on how the search engine uses MUM and BERT for safer search results.

People come to Google every day looking for solutions to protect themselves and their families. Google is continually working on new features and improvements to help you quickly locate what you need, whether it’s highlighting services in the aftermath of a natural disaster or giving time-sensitive health information.


Take a peek at how AI works. Google’s systems assist them in connecting people to vital information while avoiding potentially upsetting or harmful content, ensuring that we, as users, remain secure both online and offline.


Some key points –


  • Google will begin rolling out these changes in the coming weeks, using MUM to better recognize when a searcher is in distress.


  • BERT is a tool that Google utilizes to better recognize when a searcher is looking for explicit content.


  • According to Google, using BERT in this way has reduced “unexpected shocking results” for searchers by 30% in the last year.


 MUM is being used by Google to better serve people in personal crises.

MUM can transfer knowledge across the 75 languages it’s trained in, which can help google scale safety protections around the world much more efficiently. When they train one MUM model to perform a task — like classifying the nature of a query — it learns to do it in all the languages it knows.

For example, Google uses AI to reduce unhelpful and sometimes dangerous spam pages in search results. In the coming months,  MUM will be used to improve the quality of spam protections and expand to languages where there is very little training data. With MUM google will be able to better detect personal crisis queries all over the world, working with trusted local partners to show actionable information in several more countries.

Like any improvement to Search, these changes have and will continue to go through rigorous evaluation. 

BERT Have a better understanding of what you’re looking for.

BERT has improved google’s understanding of whether searches are truly seeking out explicit content, helping them vastly reduce your chances of encountering surprising search results.

This is a complex challenge for google to tackle for a while — but in the last year alone, this BERT improvement has reduced unexpected shocking results by 30%. It’s been especially effective in reducing explicit content for searches related to ethnicity, sexual orientation, and gender, which can disproportionately impact women and especially women of color.


Source – The Keyword