This week in AI & Machine Learning: PyTorch turns five, Meta AI announces a high-performance self-supervised algorithm, robotic curbside pickup, a new approach for a visual explanation of classifiers, and more!
Artificial Intelligence News:
PyTorch Turns Five Years Old 🥳
One of the most popular deep learning frameworks turned five years old today! PyTorch has over two thousand contributors, ninety thousand projects, and almost four million lines of “import torch” on GitHub!
To celebrate the occasion, Meta’s CTO Mike Schroepfer hosted a talk with machine learning pioneer Yann LeCun, PyTorch co-creator Soumith Chintala and Meta’s PyTorch lead, Lin Qiao.
Meta AI Announces data2vec for High-performance Self-supervised Algorithm
Meta says data2vec is the first general high-performance self-supervised algorithm for speech, vision, and text that outperforms the existing self-supervised solutions. Read about data2vec on the meta AI blog.
“We apply data2vec separately to speech, images and text and it outperformed the previous best single-purpose algorithms for computer vision and speech and it is competitive on NLP tasks.”
We created data2vec, the first general high-performance self-supervised algorithm for speech, vision, and text. When applied to different modalities, it matches or outperforms the best self-supervised algorithms. Read more and get the code:https://t.co/3x8VCwGI2x pic.twitter.com/Q9TNDg1paj
— Meta AI (@MetaAI) January 20, 2022
Technology Trend Predictions for 2022
TechRound published part two of their 2022 technology predictions, including input from over 40 experts in the field, including Carlos Anchia, Co-Founder and CEO at Plainsight. Read the entire list of 2022 tech predictions on TechRound.
“Businesses will choose to run operations leveraging cameras to optimize employee placement and assignments to mitigate exposure to physical or mental risks.”
Google Introduces StylEx, a New Approach for Visual Explanation of Classifiers
Understanding why computer vision deep neural networks make specific predictions can be a mystery. StylEx can identify individual attributes in an image that may affect the prediction probability and allow the user to adjust them individually. Read more on the Google AI blog.
Understanding how neural networks reach decisions can be challenging, but is also valuable for uses from image analysis to scientific discovery. A new approach, called StylEx, discovers and visualizes how disentangled attributes affect a classifier → https://t.co/FHSjZcw7lT pic.twitter.com/FxDJFq5eU3
— Google AI (@GoogleAI) January 18, 2022
Mitsubishi Partners with Cartken to Test Robotic Delivery
The pandemic has caused businesses to find new ways to deliver products to homes and customers who want a contactless pickup. In this pilot program, customers will have the option of robot delivery for indoor and curbside pickup at a mall. Read more about the autonomous delivery program here.
David Rolnick on Why Machine Learning is a Helpful Tool for Tackling Climate Change
David Rolnick joins one of my new favorite podcasts, The Robot Brains, to discuss why machine learning is a helpful tool for tackling climate change.