Google I/O 2022: all the news of the event

The multiple search for local information and the digital and detailed presentation of the cities in Immersive View are two of the innovations that the services of Googlein this case Finder and Maps, along with a new natural conversational language model that will allow the Assistant to understand pauses and auto-corrections.

This Tuesday the developer conference started Google I/O, an annual meeting that recovers the face-to-face public after two years due to the pandemic and that has hosted on the stage of the Shoreline amphitheater, in Mountain View, California (United States), the news of the company’s main services.

LOOK: Google adds Quechua and Aymara to its translator

These developments have to do with the way the company itself uses artificial intelligence (AI) to help people get information more efficiently. In this sense, the Search Engine will be updated at the end of the year with the multiple search close to the user.

Google Search is updated. (Google/)

This extension of the multiple search, which combines image and text search, allows the user to indicate that the results are local by typing ‘near me’ from a localized online image, for example, of a plate of food. This will allow you to get information on restaurants that serve you, or that style of cuisine, in your vicinity.

This multiple search for local information will arrive in a few months first in English, but the company has plans to extend it to other languages.

LOOK: Google strengthens its alternative messaging service to WhatsApp

Google Lens will also implement a new function, scene exploration, with which the user will be able to perform a multiple search based on what they see through their mobile camera. For example, if you point to a shelf where chocolate bars are displayed, you can see information overlaid on the items, as well as type characteristics (dark chocolate or if it has nuts) to reduce the options that are highlighted.

MAPS

Another service that will incorporate new features is Google Maps, which will introduce sustainable routes in Europe in the coming months. It will also feature immersive Viewa new view with which the user will be able to move around a high-detail digital and three-dimensional recreation of a city and even inside some establishments.

Google updates Maps. (Google/)

Immersive View performs the processing in Google cloud services, which makes it easy for any user to access this function from any mobile. It requires an internet connection to use, and will be available first, on Android and iOS, in Los Angeles, London, New York, San Francisco, and Tokyo.

Developers will have the ARCore Geospatial API at their disposal, with which they will be able to use Live View technology to create experiences in third-party applications, such as finding a parking lot, navigating locations or playing location-based games (with augmented reality). .

via GIPHY

ASSISTANT

As far as the Assistant is concerned, the work has been directed towards making conversations increasingly easier and more natural. One of the novelties is ‘Look and Talk’, or look and talk, which activates the assistant just by looking at the device, without the need to pronounce the activation command (Hey, Google and Ok, Google).

This function uses facial recognition and performs the entire process, both for this element and for voice recognition, on the device itself, which is why the company has highlighted the security and privacy of this process. It’s coming this week to Android and to Nest Hub Max devices in English in the US, but it’ll also be available for iOS in mid-May.

LOOK: Peruvians will spend more in the next online discount campaigns, why and how will they spend?

On the other hand, quick phrases, which allow the Assistant to execute actions without the activation command, such as asking for the time or turning the lights on or off, are currently available for Pixel 6 but will arrive this summer on Nest Hub Max in English in USA.

The company has also shared a new natural conversation language model, which it will roll out early next year. The novelty is that the Assistant will be able to better understand elements such as pauses, interruptions and autocorrections, and even the sound ‘Mmmm’ or ‘hum’.

Along with this novelty of Assistant, the company has also announced the second version of LaMDA, the conversational model that the company announced last year. This model has advanced to, based on a user request, offer multiple subtasks or advice, as can be seen in the AI ​​Test Kitchen app.

LOOK: Is a world without passwords possible? Apple, Google and Microsoft are already on their way to it

TRANSLATOR

Finally, the Translator will also receive news. This service is about improving the quality of translations, especially in languages ​​for which there are fewer resources. With this idea, Google has announced the incorporation of 24 new languages ​​to Translator, among which are three American indigenous languages: QuechuaGuarani and Aymara.

Google introduces Quechua and Aymara in its translator. (Google/)

With this addition, Translator now has support for 133 languages, and the 24 new additions cover more than 330 million people around the world.

The company has also explained a change in focus in the way they incorporate new languages ​​into the Translator. Until now, to train in a new language, the model had to be provided with a large number of examples.

The new approach adopted by the company is a monolingual model based on text in one language, without direct translation, instead of contrasting the same text between two languages.

OTHER SERVICES

Pixel 6 features software enhancements that allow people to be captured with a more accurate and realistic skin tone, in an attempt to combat racial bias. The company has announced a skin tone scale, developed with Dr. Ellis Monk of Harvard University (United States), which can be tested from a dedicated website.

The company has also incorporated new features in Photos and Web to move towards greater inclusivity. Some image searches will start to show filters for skin tones, for example, those referring to types of makeup. In Photos, a natural skin tone filter, based on the Monk scale, will be introduced.

The company has also announced a couple of new features for its streaming video service, the first, the arrival of auto-generated chapters, which will be applied automatically to some 80 million videos in the coming year, and the second, the extension from automatic subtitles to the mobile version.

Source: Elcomercio