This week in AI & Machine Learning: Efforts to improve Google Translate, AI to help keep the ocean clean, and more.
A Note from the Author
Factors including (but certainly not limited to) the ongoing pandemic and the rise of eCommerce have forced brick-and-mortar retailers to evolve to survive. Vision AI is helping industry leaders stand out by transforming store operations, security, and the end-to-end shopper experience. Learn more about computer vision’s role in retail’s ongoing revolution.
Artificial Intelligence News
Learning New Languages with AI’s Help
While Google Translate’s list of languages may look exhaustive, its more than 100 options represent just around 1 percent of the world’s tongues. Even this week’s addition of 24 new languages is just a small step toward making Google Translate accessible and useful for people everywhere. New research from Google addresses some of the challenges to broadening Google Translate’s scope and describes how engineers have developed a simple and practical method for zero-resource translation with monolingual datasets. Read more about Google is working toward its goal of 1000 new languages on Google Translate.
AI for Tracking Ocean Plastic to the Source
While Google Translate’s list of languages may look exhaustive, its more than 100 options represent just around 1 percent of the world’s tongues. Even this week’s addition of 24 new languages is just a small step toward making Google Translate accessible and useful for people everywhere. New research from Google addresses some of the challenges to broadening Google Translate’s scope and describes how engineers have developed a simple and practical method for zero-resource translation with monolingual datasets. Read more about Google is working toward its goal of 1000 new languages on Google Translate.
Quicker, More Efficient Training for Vision Transformers
Though vision transformers (ViTs) can support a range of computer vision use cases, AI professionals face a number of obstacles to using them as effectively as possible. Large ViTs improve the performance and accuracy of computer vision models, but they require significant resources. Training just one model could take months worth of effort and thousands of GPUs. Researchers from Meta recently shared a summary of their work to develop optimizations for training ViTs. These optimizations double the throughput in ViT training and improve accelerator utilization, saving organizations time and reducing the environmental impact of AI initiatives.
Join our Community
See you next week! Until then, keep the conversation going on Plainsight’s AI for All Slack Channel.
About the Author & Plainsight
Bennett Glace is a B2B technology content writer and cinephile from Philadelphia. He helps Plainsight in its mission to make vision AI accessible to entire enterprise teams.
Plainsight’s vision AI platform streamlines and optimizes the full computer vision lifecycle. From data annotation through deployment, customers can quickly create and successfully operationalize their own vision AI applications to solve highly diverse business challenges.