Multitask Unified Model (MUM) Algorithm in 2022
Google’s multimodal algorithm is designed to provide answers to complex questions by assessing information across multiple language text, images, video and audio.
MUM is an A.I. designed to analyze and generate content as a user would. It’s an evolution of the Google’s BERT algorithm (Bidirectional Encoder Representations: created to understand the context of a search query), except it is 1000 times more powerful.
The main goal of MUM is to address complex search queries. Currently, these types of queries require an average of eight searches to resolve. If you remember in my Google RankBrain Algorithm discussion both on my website and (Google RankBrain video on my channel), I talked about supervised and unsupervised learning algorithms.
Google Algorithm is not sophisticated enough to provide information of more complexed queries.
This means Google search engine and all other search engines can provide information which is searched before and restored in their search engines databases. Google Mum algorithm can literally run the Internet to find an answer that best suits a specific query.
Google MUM is able to provide information through text, image, video etc. only if humans have already given this information in the database. Google MUM aims at making shopping an easier task for the user. Google will be introducing a new button that will enable the user to instantly shop the image they are looking at on the screen.
Removing Language Barriers
Language can be an important barrier to accessing information. MUM has the potential to break down these boundaries by transferring knowledge across languages. It can learn from sources that aren’t written in the language a user wrote in Google’s local search engine, and help bring that information to him/her.
Google Mum’s Challenge
If a user has not searched a query before, MUM cannot provide an exact answer to that query. Because there is nothing about it in any languages in Google Search Database.
MUM is a powerful artificial intelligence algorithm and provides information as it is stored in any language in Google search engine. This can create privacy issue for individuals, companies, governments and private sectors!
I worked on well over, 7000 projects for companies in Europe, North America and South Asia. Some of these projects were related to brand reputation and reputation management. One of my cases was fixing the reputation problem of an innocent Swedish businessman who was defamed due to bullies on social media, mainly Twitter! As a result he lost his business, his marriage and had to move to a certain country to be at peace and live a quiet life.
I helped him restore his reputation online and promoted him for 2 years. It took a lot of resources and hard work along with implementing certain multilingual SEO, social media marketing and content creation strategies to fix it.
However, certain information in other languages cannot be deleted online. It takes a lot of time, more resources, human workforce and money. No matter how powerful one is in SEO and networking, it is improbable to fix it in a limited time in order to prevent the damage to his life! Online privacy is a sensitive matter!
Some information cannot be and must not be revealed online in other languages due to political and privacy reasons. This can lead to dangerous consequences for certain individuals, companies and governments!
Implementing multilingual information fetching solutions in Google search engine can be dangerous! Have your engineers at Google thought about that?! We can push down negative information per each language in local Google search, now you want to reveal it?
This is something that Google Search Engine Engineers should think about!
It saves a lot of money and resources for Google Company as well!