This week in AI & Machine Learning: DeepMind uses AI for nuclear fusion, robots in White Castle, computer trends in machine learning, and more!
Artificial Intelligence News:
DeepMind Uses Reinforcement Learning AI to Run a Nuclear Fusion Reactor
Google’s DeepMind published and EFPL (the Swiss Federal Institute of Technology Lausanne) published research showing the capabilities of using reinforcement learning to keep nuclear fusion plasma stable that could advance fusion research. Read the full paper at Nature.
Very happy to share my secret project for the last 3 years: we were busy getting AI to run a nuclear fusion reactor. Spoiler alert: it works!
— 317070 (@317070) February 16, 2022
Deep reinforcement learning is pretty great at working with sci-fi things where human intuitions tend to break down. https://t.co/r7hPzuxuOS
White Castle Expands Partnership with Miso Robotics
The Flippy 2, a robot capable of making french fries and burgers, continues to expand its presence in quick-service restaurants. White Castle is the latest chain of restaurants to increase its robotic workforce. Read more about the partnership at Meat & Poultry.
Meta Releases A New Resource to Help Measure Fairness in Speech Recognition AI
Most voice assistants still tend to be biased towards specific patterns of voices. Meta hopes the new benchmark dataset they’re releasing can increase the fairness of Speech Recognition across populations.
Read more or download the fairness dataset at the Meta AI Blog.
Compute Trends Across Three Eras of Machine Learning
Spoiler alert, computer trends are going up, almost doubling every six months! This research paper covers the compute trends across three eras of deep learning: the Pre Deep Learning Era, the Deep Learning Era, and the Large-Scale Era. Read the paper on Arxiv.
**ML training compute has been doubling every 6 months since 2010!**
— Lennart Heim (@ohlennart) February 15, 2022
Our preprint "Compute Trends Across Three Eras of Machine Learning" is out. https://t.co/XYdhY4cGw9
🧵 Thread below ↓
1/ pic.twitter.com/BCB1Y3TnVQ
Boost your Model’s Accuracy Using Self-supervised Learning with TensorFlow Similarity
TensorFlow Similarity is a python package that makes it easier to train machine learning models to detect similarity. TensorFlow Similarity now supports self-supervived machine learning to boost model accuracy on unlabeled data. Read more on the TensorFlow blog.
Why Open Source AI Contributions Are Important
Ross Wightman, a prominent contributor to the AI/ML open-source movement, joins Pieter Abbeel on The Robot Brains Podcast to discuss his journey and the importance of open source for artificial intelligence.