In an effort to give better results, the digital guru that has the answers to all our questions, Google, makes constant updates to its algorithms to adjust the way it handles search results. And every time that happens, marketers are filled with doubts that they would like to “google.”
And it is that a couple of small changes in the algorithms can make a website or the content of a brand go to heaven of the first search results or to the hell of invisibility.
But beyond the SEO campaigns that occupy brands and advertisers so much, Google is developing a matrix plan to make its search engines smarter and more conversational, in a way that offers useful and reasoned solutions, rather than generic results.
This as a way to evolve their current algorithm, BERT (Bidirectional Encoder Representations from Transformers) or Bidirectional Encoder Representation of Transformers (we leave it to BERT, for the “buddies”) is not as efficient as to answer questions in the way in which a person would do it.
What it does, like any other traditional search engine, is simply to return a series of results according to the keywords that the user enters, as well as their location and their language. They are generic answers that do not understand the real needs of the people behind their searches, nor do they consider personal preferences.
MUM: a new, smarter algorithm
For this reason, the Californian technology firm is working on what will be its new artificial intelligence algorithm: MUM, which is the acronym for Multitask Unified Model or what is the same, unified model of multiple tasks.
This model is based on artificial intelligence to give extensive and reasoned answers, joining information from different sources, since it has a neural network of a thousand nodes more than its predecessor, so it resembles the mechanism of the human synapse and is, literally, a thousand times better than BERT.
A multi-tasking intelligence
As with human intelligence, to provide more useful and complete solutions, it is necessary to understand the context from which a query is raised, addressing different aspects and points of view. In the same way, MUM is a multimodal system that is capable of understanding information in different languages, web pages, images and videos, simultaneously.
Thus, to provide answers of greater complexity, MUM uses deep learning algorithmic technology with which, rather than executing automated search orders, it is able to really understand the nature of people’s queries and their context, so it can offer responses made up of lists with links from various sites with content related to the subject, which with the previous algorithm would have required up to eight different searches.
For example, if a hiker who likes to tour different ecological reserves were to search for information about a specific destination, the new algorithm would bangladesh embassy email address only answer what he specifically asked, but it would probably give him options of the best routes, information on other nearby destinations, lodging recommendations and practical advice on the weather, the most convenient days and times to go … as an expert on the subject would answer.
In the same way, this AI understands 75 languages, and although the previous algorithms did too, MUM breaks the language barrier, since it is able to understand the interactions between words and their different meanings according to the contexts of the sentences. Thus, if someone searches for something in their native language, the algorithm would respond based on the best information found in different languages, so it not only understands languages, but generates them.
In addition, thanks to the optimization of your Visual Search you will be able to respond to more advanced queries from images and videos that you will be able to identify, interpret and search for information in this regard.
Although there is still no set date for its launch, this has been a source of doubts and fears among SEO specialists, who see this as a potential threat to their work.
However, MUM will not eliminate the need to optimize individual pages, so SEO professionals will simply have to update and adapt… once again already found a product that you want to buy, before adding it to Buy Leads the basket googlo it; You will be able to discover if in another store the price is lower,