- Descubre el Multitask United Model (MUM)
Discover the Multitask United Model (MUM)
In its translation Unique Multitasking Model, it is a search algorithm
Léelo en Español
/ 2-4 mins.
This multitasking model consists of the latest update created by Google. It is specified in the company's search algorithm and is based on artificial intelligence (AI). The improvement of the search engine is the tools that facilitate responsiveness, so its improvement is accuracy. Likewise, texts, images, videos or podcasts are used to deepen and give a better answer.
The successor to MUM is BERT, which stands for Bidirectional Encoder Representations from Transformers. This promised to improve the ability to capture human language.
On the other hand, the MUM has been an evolution of BERT, that is, the main novelty of BERT was bidirectionality, this consists of analyzing a sentence in two directions, that is, the words that are to the left and right of a word key. This will allow search engines to know the contexts in greater depth.
While MUM would be explained as multidirectionality "MUM has the potential to transform the way Google helps you with complex tasks. MUM uses the T5 text-to-text framework and is 1000 times more powerful than BERT. MUM not only understands language , but also generates it" they explain on the official Google site. Likewise, it relates the queries between the words, according to a context, something similar to the real conversations of people.
"Language can be a significant barrier to accessing information. MUM has the potential to break these boundaries by transferring knowledge between languages. You can learn from sources that are not written in the language you wrote your search in and help bring that information to you" adds Google on its updates page.
Likewise, MUM is multimodal, so its ability to search and its understanding makes it superior to BERT. In addition, it has the power to interpret content in various formats, in addition to media. That creates endless possibilities of results and complete customization for Google search users.
- Compártelo con tus amigos