Chrome’s address bar has evolved to become a powerful tool beyond being a place to type in website URLs – Google calls it the ‘omnibox’ because it’s also a search field, as well as allowing you to carry out lots of other tasks – and it’s about to get a whole lot smarter, and better able to understand what you’re looking for, thanks to machine learning.
Chrome’s omnibox will be equipped to give more precise and more relevant suggestions when you’re using Chrome, and as you use Chrome over time, the AI models behind it should improve your search suggestions thanks to upgraded ‘relevance scoring’.
Announcing the new capability in a post on the Google Chromium blog, Chrome omnibox engineering lead Justin Donnelly said he’d polled colleagues asking for ways to improve the omnibox, and “The number one answer I heard was ‘improve the scoring system.’” According to XDA Developers, this scoring system is how the omnibox interprets what the user is searching for based on their typed input.
The post also explains that this improved capability will apply to Chrome across Windows, macOS, and ChromeOS.
From static to adaptive scoring models
Donnelly added that the omnibox’s scoring system did work pretty well already, but apparently, it was pretty inflexible and static, as it was ruled by “a set of hand-built and hand-tuned formulas.” These worked well for a huge range of inputs but weren’t easy to improve or adapt in new scenarios.
He said the engineering team responsible for the innovation had been working on a machine-learning-powered scoring model which is more sensitive to different metrics (like the last time you visited a website) for a while, a process which took some time, partly due to the enormous number of searches that take place every day. Now, it looks like the improved models are ready to be rolled out.
The team found that the less you visit a particular website, the less frequently the omnibox will return that site as a suggestion when processing your search queries. It also found something even more interesting – when a user spent a short amount of time navigating a specific web page, the new model also decreased that page’s relevance score.
The model’s training data revealed a pattern of users’ behavior where they’d open a page, realize that it’s not what they were looking for, and go back to the omnibox to look for something else. Donnelly said the team wanted to incorporate this finding into their model to lower the first result’s relevance score, and if not for the new machine-learning capabilities of the model, this feature could have been missed as a helpful addition.
Onwards and upwards towards more personalized and responsive browsing
It sounds like Chromium’s engineering team took their mission pretty seriously, with Donnelly stating that his team “was driven by a sincere belief in the impact of getting this right for our users.” The results of their endeavors seem to have encouraged the team to continue working in this direction, and exploring more specialized search model versions for specific environments.
Donnelly wraps up by saying that as part of this ongoing process, the team will be observing how users’ interactions with Chrome’s omnibox change over time, so that it gets a better idea of how it can continue to improve the relevance scoring. The new model will also allow the team to collect more time-sensitive signals from user activity, and then retrain, re-evaluate, and deploy enhanced models in the future.
Overall, this sounds like a positive and exciting development, that could could deliver a more intuitive and more efficient browsing experience. Chrome’s omnibox will get better at knowing your habits and understanding what you want, and it’ll also get better at knowing what you don’t. It’s expected that this new functionality will arrive with Chrome update M124.
All of that said, you’ll probably have to live with handing over even more of your data to Google regarding your moment-to-moment online habits. If you can, and if you trust Google to handle it responsibly, then you can look forward to what sounds like a well-thought-out and innovative feature.
