Google's recent update to their search algorithm, known as BERT (Bidirectional Encoder Representations from Transformers) is bringing a real revolution in the way search engines understand queries and learn.
Unlike traditional search algorithms, which use tokenization to pinpoint relevant documents, BERT works in two directions - both left-to-right and right-to-left – to learn the context of words and how they relate to each other in a sentence. Consequently, BERT can better understand queries and produce more accurate and meaningful results.
The BERT update is the first major update to Google’s search algorithm since 2018. Google’s vice president of search and AI, Pandu Nayak said “These improvements are oriented around improving language understanding, particularly for more natural language/conversational queries, as BERT is able to help Search better understand the nuance and context of words in Searches and better match those queries with helpful results.”
Google has confirmed that the BERT algorithm applies to one in ten searches made in the US. This is expected to expand to other countries in the near future.
The new algorithm promises to look beyond plain language and interpret the context of a search query. This can help to improve clarity in search queries and results. For instance, BERT can interpret the difference between words like “help”, “assist”, “aid” and “support”, or interpret longer queries that contain multiple terms or phrases.
Google’s BERT algorithm is a major leap forward for AI research and its application to search. It promises to improve the accuracy and speed of search results using natural language processing. This, in turn, means that users will have access to more relevant, accurate and timely results. Its implementation across multiple countries is sure to make search on Google that much better for all users.