What’s more, Google is getting 15% more searches every day than it’s ever had before. How does it handle them?
The fact is that one can hardly mention any Google research papers on LSI keywords that show the stage at which LSI might be implemented in its algorithm. What is certain is that today Google uses more germany mobile database advanced natural language processing algorithms to scan the ever-expanding web. Bill Slawski clearly explains here why Google rarely uses LSI for search, citing patents as of 2017 and citing as an example that the newer Google algorithm RankBrain is based on the word embedding approach.
In fact, Google uses BERT to improve the relevance of search results to user queries, which is a neural network algorithm for natural language processing, used for paragraph ranking or understanding deep semantics in videos, which seems to be much more complex than LSI. Google has gone beyond indexing pages based on a single keyword, and you need to optimize for topics, not just for a single keyword.