Machine learning on Google Chrome’s device blocks noisy notification prompts

Google Chrome has built-in phishing detection that scans pages to see if they match known fake or malicious sites (using more than just the URL, as scammers rotate it faster than they can keep up with, and now that technology is getting better. Now Google Also says that in Chrome 102 it will use machine learning that runs entirely in the browser (without sending data back to Google or elsewhere) to help identify websites that make unsolicited consent requests for notifications and silence them before they get to them. to appear.

As Google explains, “To further improve the browsing experience, we’re also developing the way people interact with web notifications. On the one hand, page notifications help deliver updates from sites you care about; on the other, prompts can for notification consent become a nuisance. To help people browse the web with minimal interruptions, Chrome predicts when permission prompts are unlikely to be granted, and mutes those prompts. In the next release of Chrome, we’ll be launching an ML model that uses these predictions completely on the device.”

Image: Google

Image: Google

In a future release, Google plans to use the same technology to customize the Chrome toolbar in real time, with different buttons appearing like the share icons or voice search icons at times and places you’re likely to use them, without additional tracking. add those house phones. And if you prefer to choose your buttons manually, that’s okay too.

Last summer, Google announced performance improvements as part of the Chrome M92 update, reducing the time to calculate phishing rating results to 100 milliseconds from 1.8 seconds. A blog post today reports that in March it updated its machine learning model to identify 2.5 times more sites that may be phishing attacks or attempting to deliver malicious downloads. Google also says it has updated the browser-based machine learning technology behind the Journeys feature to take you on a journey through old searches so that it will now figure out if the page’s language needs to be translated.

Related Posts

Leave a Reply

Your email address will not be published.