MUM stands for Multitask Unified Model and evolves Google’s artificial intelligence algorithm, BERT (Bidirectional Encoder Representations from Transformers). MUM has 1000 more nodes in its neural network than its predecessor, which brings it much closer to the functioning of human synapses.
Originally in Spanish by Jaume Vicent in TreceBits
MUM is Google’s new Artificial Intelligence that can provide extensive and reasoned answers, extracting and linking information from different sources in more than 75 languages.
MUM can understand complex queries, i.e., requests made by users through the search engine, and return answers made up of lists of links to pages with different types of content but always inherent to what the user is asking for.
This new algorithm can provide such complex answers that the previous algorithm would have required up to eight different searches. In other words, Google MUM can give thoughtful, extensive, and complete answers to each search in the same way that a subject matter expert would. This advance means that someone searching for information will need fewer searches to find information on a particular topic in the future.
For example, imagine a hiker who, after doing a complete route in northern France, wants to do a route in the Atlas Mountains next autumn. To do so, the hiker will have to carry out several searches on Google until he finds all the information he needs. However, Google MUM will be able to understand what the user wants and provide an answer based on the comparison between the two routes, the physical preparation, and the necessary equipment.
To achieve this range of response, the AI has been trained with content in 75 different languages to understand the interactions between words and their various meanings depending on the contexts of the sentences, even with the most ambiguous ones.
In this sense, route information in the Atlas Mountains may only be in Arabic. However, MUM can “break” the language barrier and offer users the most similar content, but in their language. This is why Google says that MUM is capable of “generating language,” not just understanding it.
Moreover, MUM’s comprehension is not limited to text alone but is also capable of understanding and interpreting images, videos, and audio. This is because Google is looking to refine its search engines for visual queries and “Visual Search,” so it may be possible in the future to supplement questions with images.
MUM has brought many doubts, especially among SEO professionals who see this AI as a threat to their work. However, in a discussion on Reddit, John Mueller was asked if MUM would kill SEO. He was reassuring and pointed out that MUM will not eliminate the need for pages to optimize, just that SEO experts will have to update adapt.
