Google Search is able to understand human language with the help of several AI models that all work together to find the most relevant results.
Information about how these AI models work is explained in simple terms by Pandu Nayak, Google’s Vice President of Search, in a new article on the company’s official blog.
Nayak demystifies the following AI models, which play a major role in how Google displays search results:
- Neural matching
- my mom
None of these models work alone. They all help each other by performing various tasks to understand queries and match them to the content searchers are looking for.
Here are the key takeaways from Google’s behind-the-scenes look at what AI models do and how it all translates into better results for researchers.
Google AI models explained
Google’s first artificial intelligence system, RankBrain, was launched in 2015.
As the name suggests, RankBrain’s goal is to discover the best ranking for search results by ranking them according to their relevance.
Despite being Google’s first deep learning model, RankBrain continues to play a major role in search results today.
RankBrain helps Google understand how words in a search query relate to real-world concepts.
Nayak explains how RankBrain works:
For example, if you search for ‘what consumer address is at the highest level in a food chain’, our systems learn from seeing these words on different pages that the food chain concept may relate to animals, rather than human consumers.
By understanding these words and matching them with their related concepts, RankBrain understands that you are looking at what is commonly referred to as an ‘apex predator’.
- We see: A Complete Guide to the Google RankBrain Algorithm
Google introduced neural matching to search results in 2018.
Neural matching allows Google to understand how queries relate to pages using knowledge of broader concepts.
Instead of digging through individual keywords, neural matching scans entire queries and pages to determine which concepts they represent.
With this AI model, Google is able to cast a wider net when we scan its index for content relevant to a query.
Nayak explains how neural matching works:
“Take the research statistics on how the green is managed” for example. If a friend asked you this, you would probably be very confused. But with neural matching, we are able to make sense of it.
By looking at the broader representations of the concepts in the query—management, leadership, personality, and more—neural matching can understand that this researcher is looking for management advice based on common color-based personality evidence.”
- We see: What is Google’s Neural Matching?
BERT was first introduced in 2019 and is now used for all queries.
It’s designed to accomplish two things – retrieve relevant content and rank it.
BERT can understand how words relate to each other when used in a certain sequence, ensuring that important words are not excluded from the query.
This sophisticated understanding of language allows BERT to rank web content for relevance faster than other AI models.
Nayak shows how Bert works in practice:
For example, if you search for ‘can you get medicine for someone’s pharmacy,’ BERT understands that you’re trying to find out if you can buy medicine for someone else.
Before BERT, we took this short preposition for granted, often sharing findings on how to fill a prescription. Thanks to BERT, we understand that even small words can have big meanings.”
- We see: Google BERT Update – What does that mean
Google’s latest AI milestone in search – the Multi-Task Unified Model, or MUM, is introduced in 2021.
MUM is a thousand times more powerful than BERT, and is able to understand and generate language.
She has a more comprehensive understanding of global information and knowledge, as she is trained across 75 languages and many different tasks simultaneously.
MUM’s understanding of language extends through images, text, and more to come. This is what it means when you hear MUM referred to as “multimedia.”
Google is in the early days to realize the potential of MUM, so its use in search is limited.
Currently, MUM is being used to improve searches for COVID-19 vaccine information. In the coming months, it will be used in Google Lens as a way to search using a combination of text and images.
- We see: What is Google MUM?
Here is a summary of what the main AI systems at Google are and what they do:
- RankCategorize content by understanding how keywords relate to real-world concepts.
- Neural matching: gives Google a broader understanding of concepts, which expands the amount of content Google can search.
- Bert: Allows Google to understand how words can change the meaning of queries when used in a certain sequence.
- my momUnderstands global information and knowledge across dozens of languages and multiple modalities, such as text and images.
These AI systems work together to find and rank the most relevant content for a query as quickly as possible.
Source: The Google
Featured image: IgorGolovniov/Shutterstock