Google MUM: New technology for complex search queries

Not all searches we do on the web have a low level of complexity.  We often come across topics that require a high number of searches to find what we want to know - on average, Google considers that people do eight searches to find the answer to complex tasks.  However, in the very near future, this will no longer be necessary!  This is what Google's new tool, MUM - Multitask Unified Model, promises.  Continue reading the text and understand everything about the subject!  What is MUM?  MUM (Multitask Unified Model) is Google's new feature that will replace BERT, launched in 2019.  As mentioned above, it is not yet possible to find answers to complex issues with simple research, in the way that an expert on the subject would treat the topic.  On the contrary, several refined searches are necessary for the user to get (or get close to) the answer they need.  Want an example?  Imagine that you took a test at a certain institution to apply for a job vacancy.  However, now you will take a test from another institution, different from the first one, but with the same objective - the job vacancy.  It's normal that you want to know what to do differently to prepare yourself, after all this is a very important step in your life.  However, to research this topic on Google, you would certainly need to do several refined searches to reach the desired answer.  However, with MUM this will no longer be necessary!  According to Google, it is 1.000 times more powerful than BERT.  It is trained in 75 languages ​​and can perform different tasks at the same time, such as reading a text, understanding its full meaning, using audio and videos to study more about the subject, as well as researching in different languages ​​and producing complete content with everything related to the topic.  With it, the example above could be easily solved, providing recommendations for courses, online tests, books, and much more!  In fact, connected to Google Lens, it will even be possible to take a photo of a book, for example, and ask if it is suitable for preparing for the test in question.  It’s impressive how he humanizes the entire search process, isn’t it?!  Here's a timeline of major Google updates: December 2000: Google launches the browser "toolbar"; February 2004: Google implements a variety of changes, including massive index expansion, latent semantic indexing (LSI), and increased attention to anchor text relevance; January 2005: To combat spam and control outbound link quality, Google, Yahoo, and Microsoft launch "nofollow"; June 2005: Google launches a new version of personalized search, designed to shape future queries based on users' choices; August 2008: Suggest (later renamed Google Instant) launches, which displays suggested searches in a drop-down list below the search box as visitors type their queries; November 2010: The magnifying glass icon launches, which appears in Google search results and allows search visitors to quickly view landing pages directly from the SERPs; January 2012: Google announces changes such as image search landing page quality detection, more relevant site links, and related query improvements; January 2017: Google begins implementing a penalty to punish interstitials and aggressive pop-ups that can harm the mobile user experience; June 2017: Google officially launches its job portal, with data collected from major providers such as LinkedIn and Glassdoor; March 2018: Google announces that the mobile-first index is being rolled out, making page speed a ranking factor for mobile results; December 2019: Google confirms that BERT is being rolled out internationally, in 70 languages; May 2021: Google announces the MUM tool, which will replace BERT, but there is no release date yet.  What is the difference between MUM and BERT?  The big goal of MUM is to help users get to the answers they're looking for with less searching.  The big difference between BERT, as mentioned, is that MUM is 1.000 times more powerful.  What to expect from MUM?  BERT also aimed to lead to a better understanding of queries and content, but with the help of AI - artificial intelligence - MUM will use audiovisual tools to understand and deliver, quickly and efficiently, what the user is looking for.  Google makes it clear that the new tool will undergo several tests before going live, as it has done with other updates, paying close attention, for example, to patterns that indicate bias/prejudices in machine learning to avoid introducing these prejudices into the systems.  Therefore, there is no deadline for the launch of this long-awaited tool, but, as always, it is necessary to prepare yourself, and of course, produce quality content that always meets the needs of your target audience and your persona!  Has your company prepared itself for the increasingly intelligent and humanized updates to the platform?

Google MUM: New technology for complex search queries

Google MUM: New technology for complex search queries

Not all searches we do on the web have a low level of complexity.

We often come across topics that require a high number of searches to get to what we want to know – on average, Google considers that people do eight searches to arrive at the answer to complex tasks.

However, in the very near future, this will no longer be necessary! This is what the new tool promises. Google, the MUM – Multitask Unified Model.

Continue reading the text and understand everything about the subject!

What is MUM?

O MUM (Multitask Unified Model) is the new Google product that will replace the BERT, launched in 2019.

As mentioned above, it is not yet possible to find answers to complex issues with simple research, in the way that an expert on the subject would treat the topic.

On the contrary, several refined searches are necessary for the user to get (or get close to) the answer they need.

Want an example?

However, now you will take a test from another institution, different from the first one, but with the same objective – the job vacancy.

It's normal to want to know what to do differently to prepare yourself, after all, this is a very important step in your life. However, to research this topic on Google, you would certainly need to do several refined searches to get the answer you're looking for.

However, with MUM this will no longer be necessary! According to Google, it is 1.000 times more powerful than BERT.

He is trained in 75 languages and can perform different tasks at the same time, such as reading a text, understanding its full meaning, using audio and videos to study more about the subject, as well as researching in different languages ​​and producing complete content with everything related to the topic.

With it, the example above could be easily solved, providing recommendations for courses, online tests, books, and much more!

In fact, connected to the Google Lens, it will even be possible to take a photo of a book, for example, and ask if it is suitable for preparing for the test in question.

It’s impressive how he humanizes the entire search process, isn’t it?!

What is the difference between MUM and BERT?

The big goal of MUM is to help users get to the answers they're looking for with less searching.

The big difference between BERT, as mentioned, is that MUM is 1.000 times more powerful.

The big difference between BERT, as mentioned, is that MUM is 1.000 times more powerful.
Difference between Google's MUM and BERT

What to expect from MUM?

BERT also aimed to lead to a better understanding of queries and content, but with the help of AI – artificial intelligence – MUM will use audiovisual tools to understand and deliver, quickly and efficiently, what the user is looking for.

Google makes it clear that the new tool will undergo several tests before going live, as it has done with other updates, paying close attention, for example, to patterns that indicate bias/prejudices in machine learning to avoid introducing these prejudices into the systems.

Therefore, there is no deadline for the launch of this long-awaited tool, but, as always, it is necessary to prepare, and of course, produce quality content that always meets the needs of your target audience and your company. person!

Has your company prepared itself for the increasingly intelligent and humanized updates to the platform?

Count on GS2 Marketing Digital for this mission!

ebook wants to learn everything about marketing automation

Leave the first comment

Contents