Using AI to bring children’s drawings to life
Meta AI researchers are working to overcome this challenge so that AI systems will be better able to recognize drawings of human figures in the wildly varied ways that children create them. Kids’ drawings are often constructed in abstract, fanciful ways, so if a figure’s feet are placed precariously or if both arms are on the same side of its body, it confuses even state-of-the-art AI systems.
A communication tool for people with speech impairments
Project Relate is a new Android app that aims to help people with speech impairments communicate more easily with others and interact with the Google Assistant. The app transcribes your speech to text in real time, so you can copy-paste text into other apps, or let people read what you want to tell them. We are now looking for English-speaking testers in Australia, Canada, New Zealand and the United States.
Harmful content can evolve quickly. Our new AI system adapts to tackle it.
Few-Shot Learner (FSL) is a new AI technology that can adapt to take action on new or evolving types of harmful content within weeks instead of months. It works in more than 100 languages, and it also learns from different kinds of data, such as images and text. Unlike previous systems that relied on pattern-matching with labeled data, FSL is pretrained on general language, as well as policy-violating and borderline content.
How AI is making information more useful
Advances in artificial intelligence will transform the way we use that information, with the ability to uncover new insights that can help us both in our daily lives and in the ways we are able to tackle complex global challenges. In the coming months, we’ll introduce a new way to search visually, with an ability to ask questions about what you see in the images.
XLS-R: Self-supervised speech processing for 128 languages
XLS-R is based on wav2vec 2.0, our approach to self-supervised learning of speech representations. It’s based on 436,000 hours of publicly available speech recordings. XLSR-53 was released last year, covering nearly two and a half times more languages than its predecessor. We found that our largest model, containing over 2 billion parameters, performs much better than smaller models, since more parameters are critical.
How Abigail Annkah is using AI to improve maps in Africa
I quickly developed an interest in using data-driven approaches to solving pressing societal challenges, leading me to work on biochemical image segmentation for my master’s thesis. I then joined the Google AI center in Ghana as an AI resident and after two years gained a full-time role as a research software engineer. The Open Buildings project uses AI to provide a digital footprint of building locations and geometry across most of Africa.
Reducing city transport emissions with Maps and AI
EIE is helping cities and communities reduce carbon emissions by 2030. It will be one of the largest ever collections of high-quality data sources in the world. EIE aims to help more than 20,000 cities and more than 1,000 people a year to reduce emissions. It’s a global effort to reduce greenhouse gas emissions by reducing carbon emissions in cities and cities.
These researchers are bringing AI to farmers
Both Diana Akrong and Courtney Heldreth are part of Google’s People + Artificial Intelligence Research (PAIR) group. They are working on how AI can help better the lives of small, local farming communities in the Global South. They and their team want to understand what farmers need, their practices, value systems, and what their social lives are like.
OpenAI’s API Now Available with No Waitlist
Starting today, developers in supported countries can sign up and start experimenting with our API right away. Improvements to our API over the past year include the Instruct Series models that adhere better to human instructions, specialized endpoints for more truthful question-answering, and a free content filter to help developers mitigate abuse. As our systems evolve, we expect to continue streamlining the process for developers, refining our Usage Guidelines, and allowing even more use cases over time.
Detecting manipulated images: The Image Similarity Challenge results and winners
NeurIPS 2021 is sharing results of the Image Similarity Challenge, an online AI research competition to create systems that can assess the similarity of two images in order to identify manipulated images. The competition drew more than 200 participants, who trained and tested models using a first-of-its kind dataset. The top-scoring teams, VisionForce and Iyakaap, each won $50,000.
Partnering with the NSF on a research institute for AI to improve care for older adults
U.S. government-funded R&D has yielded remarkable progress for society, and today it is an important engine for AI research. Last year, we were proud to announce our partnership with the National Science Foundation to provide $5M to support the establishment of national research institutes working in the area of Human-AI Interaction and Collaboration (HAIC) This partnership is part of a more than $300M NSF investment in AI Research Institutes.
GPT-3 Powers the Next Generation of Apps
Over 300 apps are using GPT-3 across varying categories and industries, from productivity and education to creativity and games. Viable identifies themes, emotions, and sentiment from surveys, help desk tickets, live chat logs, reviews and more. Fable Studio is creating a new genre of interactive stories using the API to help power their story-driven “Virtual Beings”
Expanding our ML-based flood forecasting
In 2018, Google began its flood forecasting initiative to help combat the catastrophic damage from floods each year by equipping those in harm’s way with accurate and detailed alerts. This work is a part of Google’s broader Crisis Response program which provides people access to trusted information and resources in critical moments. In 2021, our operational systems were further expanded to cover an area with over 360 million people.
OpenAI Codex
OpenAI Codex is the model that powers GitHub Copilot, which we built and launched in partnership with GitHub a month ago. Codex can interpret simple commands in natural language and execute them on the user’s behalf. It has a memory of 14KB for Python code, compared to GPT-3 which has only 4KB. Codex is most capable in Python, but it is also proficient in over a dozen languages including JavaScript, Go, Perl, PHP, Ruby, Swift and TypeScript.
Introducing the AI Research SuperCluster — Meta’s cutting-edge AI supercomputer for AI research
Meta has designed and built the AI Research SuperCluster (RSC) It will be the fastest AI supercomputer in the world when it’s fully built out in mid-2022. RSC will help Meta’s AI researchers build new and better AI models that can work across hundreds of different languages; seamlessly analyze text, images, and video together; develop new augmented reality tools.
Solving Math Word Problems
We’ve trained a system that solves grade school math problems with nearly twice the accuracy of a fine-tuned GPT-3 model. It solves about 90% as many problems as real kids: a small sample of 9-12 year olds scored 60% on a test from our dataset. This is important because today’s AI is still quite weak at commonsense multistep reasoning, which is easy for grade school kids.
This archaeologist fights tomb raiders with Google Earth
Dr. Gino Caspari studies the ancient Scythians with the Swiss National Science Foundation. More than 90% of the tombs have already been destroyed by raiders looking to profit off what they find. To track his progress, he began mapping these burial sites using Google Earth. The work isn’t easy, from dealing with extreme temperatures, to swamps covered with mosquitos.