This week - real-world challenges for AGI; scientists reversed blindness with CRISPR; what neurons and snowflakes can teach us about AI; and more!
More Than A Human
Most powered exoskeletons turn their wearer into a little forklift, restricting them to lifting weights in a vertical plane. This exoskeleton, however, was designed to also assist in rotating and twisting movements such as throwing a heavy bag.
AI acquired a new skill - accurately predicting the shape of proteins. Thanks to AlphaFold and RoseTTAFold, researchers got new powerful tools with which they can create new drugs and better understand how life itself works.
AI researchers created a machine learning model to predict new drugs before they enter the market. This model can help law enforcement agencies cut identification time down from months to days, and speed up the process of regulating new versions of psychoactive drugs.
The fact that no two are alike could be the key to designing AIs that can adapt to change, argues Daniel Goodman. To back this up, his team recently showed that introducing variation by slightly tweaking every artificial neuron in an AI system improves the accuracy of a simulated neural network’s learning performance by up to 20%.
"Advances in AGI research will supercharge society’s ability to tackle and manage climate change - not least because of its urgency but also due to its complex and multifaceted nature", writes Koray Kavukcuoglu from DeepMind.
In this video, ColdFusion briefly explains what is CRISPR and then goes into explaining how this technology has been used to restore eyesight by correcting a genetic defect in the patient's eye.