Google announced a set of new and updated security features for Chrome, almost all of which are based on machine learning (ML) models, along with some new ML-based features that aim to make web browsing a little easier. , including a new feature that will suppress notification permission requests when your algorithm thinks you are unlikely to accept them.
Starting with the next version of Chrome, Google will introduce a new ML model that will block many of these notification permission requests.
And Google to explain:
Safe browsing in Chrome helps protect billions of devices every day by displaying alerts when people try to browse dangerous sites or download dangerous files (see the big red example below). Starting in March of this year, we launched a new ML model that identifies 2.5 times more potentially malicious sites and fishing attacks than the previous model, resulting in a more secure website.
To further enhance your browsing experience, we’re also evolving the way people interact with web notifications. On the one hand, page notifications help deliver updates to sites that matter to you *; on the other hand, notifying permission requests can be a nuisance. To help people navigate the web with minimal disruption, Chrome predicts when permission requests are unlikely to be granted based on how the user has previously interacted with similar permission requests and silences unwanted guests. In the next version of Chrome, we are running an ML model that makes these predictions completely on your device.
With the next version of Chrome, this is what you’ll see if a fishing attempt is detected (left) and Chrome will silently show permission requests when the user is unlikely to grant them (right).
In a future release, Google plans to use the same technology to adjust the Chrome Toolbar in real time, with different buttons such as sharing or voice search icons appearing when and where you’re likely to use them. :
You may enjoy reading news articles in the morning (phone in one hand, cereal spoon in the other), so you share many Chrome links. Or maybe voice search is more yours, as you ask some questions on your commute to work. Either way, we want to make sure Chrome finds you where you are, so in the near future we’ll use ML to adjust the toolbar in real time, highlighting the most useful action right now (e.g. , a link to share, search by voice, etc.). Of course, you can also customize it manually.
The Chrome to Android toolbar will be tailored to your needs.
As for other new features based on machine learning, Chrome also gets a new language identification model that better determines which language a particular page is in and whether it should be translated accordingly. your personal preferences:
Earlier this year, we launched Journeys to help people get back on track online. For example, you can spend weeks planning a visit to a national park, researching attractions, comparing flights, and buying equipment. With ML and Journeys, Chrome picks up pages you’ve visited on a specific topic and allows you to easily resume where you left off (instead of scrolling through your browser history).
When you return to these hiking boots and camping guides, we also use ML to make these websites available in the language of your choice. In particular, we’ve posted an updated language ID template to determine the language of the page and whether it should be translated to match your preferences. As a result, we see tens of millions of successful translations every day.
Chrome’s Journeys feature groups your search history by topic or intent
The Chrome team says that their goal is to create a browser that is really useful and continuous, and we are excited about the possibilities that ML offers.
Last summer, Google announced performance enhancements as part of the Chrome M92 update, reducing the time it takes to calculate 100% millisecond fishing rankings in 1.8 seconds. Then Google explained:
Each time you access a new page, Chrome evaluates a collection of signals on the page to see if they match those of the fishing sites. To do this, we compare the color profile of the visited page – that is, the range and frequency of colors present on the page – with the color profiles of the current pages. For example in the image below, we can see that the colors are mostly orange, followed by green and then a touch of violet.
If the site matches a known credential fishing site, Chrome warns you to protect your personal information and prevent you from exposing your credentials.
What to see if a fishing attempt is detected
To preserve your privacy by default, Chrome’s secure browsing mode never sends images out of your browser. While this is ideal for privacy, it means that your machine has to do all the work to analyze the image.
Image processing can often result in large workloads because image analysis requires evaluating each pixel in what is commonly referred to as the “pixel loop.” Some modern monitors display more than 14 million pixels, so even simple operations on each of these pixels can result in high CPU usage *! For phishing detection, the operation that takes place in each pixel is the count of its base colors (…).
Starting with M92, Chrome now runs an image-based fishing rating up to 50 times faster at the 50th percentile and 2.5 times faster at the 99th percentile. On average, users will get their fishing ranking results after 100 milliseconds, instead of 1.8 seconds.
This benefits you in two ways when you use Chrome. First, using less CPU time to do the same job improves overall performance. Less CPU time means less battery consumption and less time with rotating fans.
Second, faster results mean Chrome can notify you sooner. Optimization reduced the percentage of requests that took more than 5 seconds to process from 16.25% to less than 1.6%. This speed improvement makes a real difference in terms of security, especially when it comes to preventing you from entering your password in a malicious place.
A blog post reports that in March, the Google Chrome team updated the machine learning model to identify 2.5 times more sites that could be phishing attacks or trying to deliver malicious downloads.
Source: Google Chrome
What do you think of the new and improved features that Google Chrome has introduced?
Do you find any useful? Which is?
Are there any that you find redundant? Which is?
In general, what do you think of Chrome today?
What browser are you working on? Which browser are you using for your personal browsing?
Chrome will turn off features such as alert () in cross-source frameworks, but developers regret that such a big change occurs without a thorough discussion
Google Chrome will no longer show secure website flags as the company continues to strive for a 100% HTTPS website
Google Chrome 100 is available with new Chrome DevTools features and a new dedicated window location management API