At Google I/O earlier last year, Google announced that it was internally exploring a new technology called MUM (Multitasking Unified Model) to help its ranking system better understand language.
Billed as a “new AI milestone in understanding information,” MUM aims to make it easier for Google to meet complex demands in search.
Google’s commitment to MUM 1,000 times More powerful than its NLP transfer learning predecessor, BERT.
It uses a T5Text-To-Text Transfer Transformer, which reframes NLP tasks into a unified text-to-text format and develops a more comprehensive understanding of knowledge and information.
According to Google, MUM can be applied to classification tasks such as document summarization, question answering, and sentiment analysis.
Clearly, MUM is a major priority within the Googleplex – what’s important to the search team should also get better attention on the SEO industry’s radar.
But is it a ranking factor in Google’s search algorithm?
Statement: Mom as a Ranking Factor
Many people have read the news about MUM when it was first revealed Naturally want to know how it affects search rankings (especially their own).
Google makes thousands of updates to its ranking algorithm every year, and while the vast majority go unnoticed, some are impactful.
BERT is one such example.
Launched globally 2019which was hailed by Google itself as the most important update in five years.
Sure enough, BERT affected approximately 10% search query.
RankBrainLaunched in Spring 2015, it is another example of an algorithmic update that has a significant impact on the SERPs.
Now that Google is talking about MUM, it’s clear that SEO professionals and the clients they serve should take note.
Roger Monty recently Wrote an article on patents He thought it could provide more insight into the inner workings of MUM.
It’s an interesting book if you want to see what’s under the hood.
For now, let’s just consider whether MUM is a ranking factor.
Evidence of MUM as a ranking factor
When RankBrain launched, it wasn’t announced until about six months later. Most updates are not announced or confirmed at all.
However, Google had done better before the impactful update happened.
For example, BERT is first announced Launched for English queries in November 2018 October 2019and launched globally later that year December.
We have more time to prepare Page Experience Signals and Core Web Vitals, which have been announced more than a year in advance the final launch of June 2021.
Google has already said that MUM is coming, and it’s going to be a big deal.
But will MUM be responsible for the ranking drops experienced by many sites in the spring and summer of 2021?
Evidence against MUM as a ranking factor
In a May 2021 presentation on MUM, Google Fellow and VP of Search Pandu Nayak, make it clear This technique doesn’t work. Not yet, anyway:
“Today’s search engines aren’t sophisticated enough to answer questions like an expert. But with a new technology called the Multitasking Unified Model (MUM), we’re getting closer to helping you with these types of complex needs. So future, you will need less searching to get the job done. “
The timeline given at the time as to when MUM-supported features and updates would go live was “in the coming months and years.”
When asked if the industry would take notice when MUM goes live in search, Google search liaison Danny Sullivan said yes.
Yes, just like BERT, I’m sure we’ll let everyone know. We will not be silent about MUM.
— Danny Sullivan (@dannysullivan) May 20, 2021
More recently, Nayak explained How Google is using AI in search and wrote,
“While we are still in the early stages of unlocking the potential of MUM, we are already using it to improve the search for COVID-19 vaccine information, and we will provide more intuitive ways to search using a combination of text and images in the next few month in Google Lens.
These are very specialized applications – so MUM is not currently used to help rank and improve the quality of search results Just like RankBrain, Neural Matching, and BERT systems. “
He also added that any future application of MUM will go through a rigorous evaluation process, including special attention to the responsible use of artificial intelligence.
MUM as a ranking factor: our verdict
Bottom line: Google does not use MUM as a search ranking signal. It is a language AI model based on Google’s open source neural network architecture Transformer.
Google will train moms as BERT does on large datasets, and then fine-tune it for specific applications on smaller datasets. That’s the test MUM uses to improve vaccine search results.
Google has mentioned specific ways it might use in the (near) future, including:
- Insight emerges based on its deep knowledge of the world.
- Presents useful subtopics for deeper exploration.
- Break down language barriers by transferring knowledge across languages.
- At the same time understand information from different formats, such as web pages, pictures, etc.
How would you optimize for MUM?
It remains to be seen.
One thing’s for sure: the intelligence of Google’s search is growing by leaps and bounds.
As Google’s search algorithms become more sophisticated and better able to determine the intent and nuances of language, attempts to deceive and manipulate will become less effective (and possibly easier to spot).
With upcoming NLP technologies 1000x more powerful than RankBrain, optimizing the human experience is more important than ever.
If you want to get ahead of MUM, focus on What the content you’re creating means It is designed to meet people’s needs.
These machines are getting closer to fully and completely experiencing the content as your intended reader/viewer.
Featured image: Paulo Bobita/Search Engine Magazine
