BERT

What is BERT?

BERT revolutionizes how search engines interpret and understand the context of words in search queries. Unlike traditional models that read text in a linear way, BERT reads text bidirectionally, meaning it looks at the context from both the left and the right side of a word. This allows search engines to better understand the intent behind a search query, leading to more accurate and relevant search results. For instance, BERT can differentiate between the meanings of the word 'bank' in the sentences 'He sat by the river bank' and 'She went to the bank to deposit money'. By grasping these nuances, BERT helps to provide searchers with results that are more closely aligned with what they are actually looking for. This advancement is particularly beneficial for long-tail queries, which are often more complex and conversational in nature.

BERT (Bidirectional Encoder Representations from Transformers) is a natural language processing model developed by Google that helps machines understand the context of words in search queries.

Examples

A user searches for '2019 brazil traveler to usa need a visa'. Before BERT, Google might have returned results about U.S. citizens traveling to Brazil. With BERT, Google understands the searcher is a Brazilian traveling to the U.S. and provides accurate visa information.

Someone searches for 'can you get medicine for someone pharmacy'. Pre-BERT, Google might have struggled with this query's context. With BERT, Google understands the searcher wants to know if they can pick up a prescription for someone else from the pharmacy.

Additional Information

BERT was introduced by Google in October 2019 and has since significantly improved search query understanding.

BERT is open-source, making it available for developers and researchers to use and improve upon in various natural language processing tasks.

References

How BERT Affects SEO and How You Can Optimize For It

BERT Explained: What You Need to Know About Google's New Algorithm

FAQ: All about the BERT algorithm in Google search - Search Engine Land