Google I/O 2022 Wonderful Review: "Knowledge and Computing" in Focus | Practice "Technology Makes Life Better for All"

Technical editor: MissD丨From the editorial department of Sifu

At 1:00 am Beijing time on May 12th, Google held the Google I/O 2022 conference at the Shoreline Amphitheatre. As this year's first annual technology event for developers, partners and Google users around the world, this event returns to offline for the first time in three years, and witnesses the excitement with users from all over the world online.

This year's conference is also the first time in three years that Google CEO Sundar Pichai personally attends and presides over the event. During the keynote, Google brought an update to its next-generation operating system, Android 13, and officially launched the Pixel 6A smartphone, Pixel Watch, and Pixel Buds Pro earphones. (Also, teasers for the Pixel tablet and Pixel 7 and Pixel 7 Pro phones were revealed ahead of time, the latter slated for a fall launch.)

At the meeting, Google also emphasized smart features such as instant translation and Flutter 3.0, and gave a quick preview of its AR glasses under development. In addition, there are many new technologies and new highlights that surprise developers and Google users, so let's review them soon!

Advance knowledge and computation: Translation, Maps

At the beginning of the conference, we saw a sharing of "Latest progress of Google I/O - advancing knowledge and computing" brought by Google CEO Sundar Pichai, and how these new technology advancements help more people communicate without barriers, make life better for everyone.

Google Translation: As we all know, translation has always been a technical challenge for languages, because the texts used to train translation models are often bilingual texts. Google Translate has added 24 new languages, with machine learning, allowing translation models to directly learn to translate a new language without ever seeing a literal translation of the new language.

Google Maps: Based on 3D mapping and machine learning technology, using computer vision and neural network technology to achieve high-quality imagery and clear building types, Google has increased the number of regional buildings on Google Maps since July 2020. times, more than 20% on Google Maps worldwide.

YouTube videos can automatically generate chapters and transcriptions: Apply DeepMind's multi-modal technology to automatically generate chapters for text, audio, and video simultaneously with greater accuracy and speed; transcribe videos by using speech recognition models (all Android and iOS users can use video transcription). Based on this technology, Google's next goal is to increase the number of videos with automatically generated chapters 10 times, from the current 8 million to 80 million next year.

Auto-subtitles for YouTube videos: This time Google is bringing auto-subtitles for YouTube videos to mobile phones. That said, users can now automatically translate video subtitles in 16 languages, allowing creators to expand global audiences.

Google Workspace: Google Docs summarises text through one of its machine learning models, Google Docs will automatically parse words and extract key points, also marking a big leap in natural language processing. Google is currently releasing summaries for other products in Workspace and will appear on Google Chat over the next few months, providing a useful summary of chat conversations, allowing users to jump directly into group chats or review key highlights.

Visual improvements on Google Meet: After Google introduced Project Starline at I/O last year, it has been testing feedback for continuous improvements to continuously improve audio and video quality in the "virtual space". And Starline-inspired machine learning supports image processing to automatically improve image quality in Google Meet.

The Next Frontier of Computing: Augmented Reality/AR

At this conference, Google also revealed the progress and plans of AR technology - Augmented Reality/AR, a new computing frontier in the future, which may further expand all. "Google is designed for the real world in a way that doesn't take you out of it. AR gives us new ways to do that."

At present, Google has invested heavily in this area and has been applying AR technology to many Google products, such as Google Lens, search, scene exploration, and real-time and immersive views in maps.

Although there is still a long way to go in the future, Google is optimistic that "AR technology can break the language barrier", understand and be understood, which is the meaning of Google's focus on knowledge and computing today.

A broader vision: Multisearch, Monk Skin Tone Scale

In the sharing of Google Search improvements brought by Google Senior Vice President Prabhakar Raghavan, we brought us Google's new vision for "Search reimagined. Any way and anywhere" this year - how to help people search in a more intuitive way Find useful information in .

Multisearch "scene recognition": Through this function, when we encounter all the indescribable things in our daily life, we can directly search in the Google application in the form of photos and questions.

This time Google has also brought new functions such as "Multisearch Near Me" and "Scenario Exploration", which is more convenient to provide the search needs of "thinking about what you think" in our daily life scenarios.

Monk Skin Tones: At this conference, Google also brought the "Monk Skin Tones Scale", a new free and open source development tool provided by Google AI, a collaboration between Google and Harvard University Dr. Ellis Monk Tools designed to take accurate photos of people with darker skin tones for building more inclusive AI experiences. Google has now officially opened the source code of Monk Skin Tones, intending to cooperate with the industry to continuously improve and improve the technology in related fields.

Google Assistant Big Improvement: Look and Talk

At present, 700 million users around the world use Google Assistant. Today, Google brings 2 new technologies-Look and Talk, which can implement the assessment of users' "mind", so that you don't have to say "Hey Google!" twice every time. , allowing people to interact with their devices in a more natural, conversation-like way.

LaMDA 2 model: This is Google's most advanced instant conversational AI to date, generating language models for conversational applications that can have conversations about any topic, to achieve "conversation and natural language processing is the way to make everyone more A powerful way to use computers easily" vision. Currently, Google has opened LaMDA to thousands of Google users, and under continuous improvement, LaMDA has greatly improved the quality and reduced inaccurate or offensive responses.

AI Test Kitchen: This function allows more people to understand the functions of artificial intelligence, and helps create lists easily through the three steps of "Imagine It", "TalkAbout It" and "List It". Also, use the AI Test Kitchen to do the things you care about with LaMDA.

To explore other aspects of natural language processing and artificial intelligence, Google also announced a new model — the Pathways Language Model, or PaLM for short. This is Google's largest model to date, trained with 540 billion parameters. PaLM has shown breakthrough performance on many natural language processing tasks, such as generating code from text, answering math word problems, and even explaining jokes.

8 TPU v4 pods: the world's largest, public machine learning center

As Google continues to innovate in infrastructure, Google has made a lot of progress in many areas in recent years. Today, Google also announced plans to invest $9.5 billion in data centers and offices across the U.S. and announced the launch of the world's largest, publicly available machine learning with eight TPU v4 pods in Oklahoma, U.S., for Google Cloud customers Center to drive innovation in fields ranging from medicine to logistics, sustainability and more.

It is worth emphasizing that this machine learning center is already running on 90% carbon-free energy, and in the upcoming 2030, it will become the first major company in the world to operate all data centers and campuses on 24/7 carbon-free energy.

Multi-device world under Android collaboration

At this I/O conference, Google also officially updated the software and hardware of its smartphones, watches and tablet devices, in addition to the Android 13 operating system, it also includes Pixel 6a, Pixel Buds Pro, Google Pixel Watch, Pixel 7 and Pixel Tablet PCs to help all of the user's devices work better together.

Android 13 operating system upgrade

Android 13 Beta 2: The highly anticipated Android 13 has been tested for a while, and although the official version has not been announced yet, at this conference, Google released the Android 13 Beta 2 version to create a "Android multi-device world" "New Vision.

Android 13 improvements: Expanded Material You design style, improved privacy and security issues with RCS (converged communications) support for Google Messages, and tested group chats with end-to-end encryption. In addition, Android 13 also updates the "limitation of personal data and phone functions that applications can use by default" and the "photo picker".

Google's new hardware surprise debut

Google Pixel Buds Pro Headphones: These are the first headphones to combine Active Noise Cancellation and Google technology for better isolation and transparency (noise reduction). There will be a small update (adding spatial audio) after release, which can be tracked via Android's Find My device. The headset is priced at $199 and will be available for pre-order on July 21.

Google Pixel 6A, Pixel 7, Pixel 7 Pro: At this conference, Google released the $449 Google Pixel 6 entry-level phone, which will be available for pre-order on July 21. In addition, Google also revealed the Pixel 7 series, which has a recyclable aluminum camera array and a glass back design, as well as an updated version based on an artificial intelligence-optimized domestic mobile Tensor processor, which will be launched this year. Android 13 will be officially launched in the fall.

Pixel Watch, Pixel Tablet: Google's new Pixel Watch watch, Pixel Tablet device featuring an improved Wear OS interface and a curved glass screen. The new device will support Google Wallet and support broad compatibility with Fitbit and is expected to be available in the fall. The tablet, which complements the Pixel phone, is expected to launch in 2023.

Other Tech Update Highlights

Flutter 3.0: Today Google officially launched Flutter 3.0, completing its roadmap from a mobile-centric to a multi-platform framework, offering support for macOS and Linux desktop apps, as well as improvements to Firebase integration, new productivity and performance features , and supports Apple Silicon. In addition, Flutter 3.0 further improves the performance related to opacity animation, supports ProMotion refresh rate on iOS such as iPhone 13 Pro and iPad Pro, and also supports foldable mobile devices.

Google Wallet: Google also launched Google Wallet at this conference, which can be used directly as a driver's license, bank card, boarding pass, and ticket, and also supports WearOS users.

As a product that competes with companies such as Apple Wallet and Samsung Pay, I believe Google Wallet will have a different surprise experience.

Overview

In the past period of time, advanced technology has not only helped mankind to overcome new challenges, but also provided assistance to global technological problems. Remember last year's Google I/O 2021 conference, where Google focused on how new breakthroughs in some of the most technically challenging areas of computer science can make Google products more helpful at critical times. At this year's Google I/O 2022, Google will bring new technologies to the Microsoft Bulid Conference and Apple's WWDC before the heavy "first launch", which can be said to be well-prepared and aggressive.

At this conference, Google joined hands with a series of new technologies such as software and hardware, as well as AI and other major upgrades, which can be said to have brought exciting momentum to the next steps of global developers. The innovation brought about by these new technologies also just confirms the "advancing knowledge and computing" and the broad vision emphasized at this Google I/O conference - "Let more people communicate without barriers, and let more people have fairness. The arena where technology makes life better for everyone and continues to work hard to create the future.”

Click on the link to see the highlights of Google I/O 2022~