MUM stands for Uniform multitask model and it is the further development of Google’s artificial intelligence algorithm, BERT (Bidirectional Encoder Representations from Transformers). MUM has 1,000 more nodes in its neural network than its predecessor, which brings it much closer to how the human synapse works.
MUM is Google’s new artificial intelligence capable of providing comprehensive and reasoned answers and of extracting and merging information from various sources in more than 75 languages.
MUM understands the “inquiries”, that is, the Requests from users by the search engine, complex and repetitive answers, consisting of lists of links to pages with content of a different kind, but which are always inherent to the needs of the user.
This new algorithm is able to provide answers that are so complex that the previous algorithm would have needed up to eight different searches. In other words, Google MUM is capable of giving thoughtful, comprehensive, and complete answers to any search, just as an expert on a subject would. This advancement means that in the future someone looking for information will need fewer searches to find information on a particular topic.
For example, imagine a hiker who, after a complete route through northern France, wants to take a route through the Atlas Mountains next fall. To do this, the hiker has to carry out various 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 both routes, physical preparation and equipment required.
In order to achieve this reaction range, the AI was trained with content in 75 different languagesso that he is able to understand the interactions between words and their different meanings according to the context of the sentences, even with the most ambiguous.
With this in mind, it is possible that the information about the routes in the Atlas Mountains is only in Arabic. However, MUM is able to “break through” the language barrier and offer the user the most similar content, but in his own language. Because of this property, Google points out that MUM is able to «Generate speech«Not just to understand.
In addition, the understanding of MUM not only remains in the text, but is able to understand and interpret Pictures, videos and audios. This is because Google wants to refine its search engines for visual search queries and “visual search” so that in the future it may be possible to supplement search queries with images.
MUM has raised a lot of doubts, especially among SEO professionals who see this AI as a threat to their work. In a discussion on Reddit, John Mueller was asked if MUM would kill SEO
