Searching for information online sometimes takes creativityâ€”depending on how you word a query, Google can bring back some wonky results. Thatâ€™s why Google announced today that itâ€™s improved search to better understand natural language for â€œqueries [it] canâ€™t anticipate.â€
Part of the issue is that we imperfect humans donâ€™t always know how to spell what weâ€™re looking for, or even word it in a way that makes sense to computers. If you phrase a query conversationally, you run the risk of getting garbage results back. Thatâ€™s why itâ€™s way more common to use what Google calls â€œkeyword-eseâ€, where youâ€™re just typing related words together. Sort of the equivalent of baby talk, for a computer algorithm.
To combat this, Google says itâ€™s using an open-sourced neural network to train search for better natural language processing. Itâ€™s called Bidirectional Encoder Representations from Transformers (BERT), and in a nutshell, lets search algorithms consider linguistic context instead of relying solely on keywords. In particular, prepositions like â€œforâ€ and â€œtoâ€ can trip up a regular search. One example that Google provided was the search results for the query â€œ2019 brazil traveller to usa need a visa.â€ Before applying BERT, the results would give backlinks about U.S. citizens travelling to Brazil.
Googleâ€™s seen success with its natural language processing in recent years with Google Assistant. Of the three major voice assistantsâ€”Siri and Alexa being the other twoâ€”Assistant is the best at understanding and answering conversational commands and queries. Itâ€™s of course, not perfect, but as someone with all three in my apartment, I donâ€™t have to think quite as hard about the exact phrasing I use when speaking to my Google Nest Hub.
According to Google, using BERT will help improve one-in-ten English-language searches in the U.S. Thatâ€™s not to say that BERT wonâ€™t roll out to other languages sometime down the line. Google says its taking improvements from English models and is currently seeing significant improvements in Korean, Hindi, and Portuguese.