News

Self-Driving Truck Encounters the Unexpected, Then Reaches a Milestone

Starsky Robotics engineers tested and developed the autonomous-driving system, which has been installed in the Freightliner Cascadia for more than a year, and had spent a full week conducting dry runs on the stretch of County Road 833 in southern Florida, just north of the Everglades. They had plotted every conceivable contingency and every imaginable edge-case scenario.

Nuclear fusion on brink of being realised, say MIT scientists

Why humans learn faster than AI for now

In 2013, DeepMind Technologies, then a little-known company, published a groundbreaking paper showing how a neural network could learn to play 1980s video games the way humans do by looking at the screen. These networks then went on to thrash the best human players.

Heredity Beyond the Gene

The idea that genes encode all the heritable features of living things has been a fundamental tenet of genetics and evolutionary biology for many years, but this assumption has always coexisted uncomfortably with the messy findings of empirical research. The complications have multiplied exponentially in recent years under the weight of new discoveries.

Google Just Indexed Millions of ‘Life Magazine’ Photos Using Artificial Intelligence

Open Sourcing the Hunt for Exoplanets

Recently, we discovered two exoplanets by training a neural network to analyze data from NASA’s Kepler space telescope and accurately identify the most promising planet signals. And while this was only an initial analysis of ~700 stars, we consider this a successful proof-of-concept for using machine learning to discover exoplanets, and more generally another example of using machine learning to make meaningful gains in a variety of scientific disciplines (e.g. healthcare, quantum chemistry, and fusion research)

Reptile: A Scalable Meta-Learning Algorithm

We’ve developed a simple meta-learning algorithm called Reptile which works by repeatedly sampling a task, performing stochastic gradient descent on it, and updating the initial parameters towards the final parameters learned on that task. This method performs as well as MAML, a broadly applicable meta-learning algorithm, while being simpler to implement and more computationally efficient.

Researchers Can ‘See’ Your Memories While You Sleep

New research published today by Cairney and his colleagues in Current Biology has made significant headway on this issue by linking sleep spindles—spontaneous bursts of brain activity—to memory processing in a sleeping brain. Incredibly, the researchers claim it is possible to determine the content of the memory being processed by analyzing this brain activity.

Can retrocausality solve the puzzle of action-at-a-distance?

Einstein’s objections to quantum mechanics began very early. Schrödinger’s version of the theory introduced a new mathematical entity, the wave function, which seemed to allow the position of an unmeasured particle to be spread out across an arbitrarily large region of space. When the particle’s position was measured, the wave function was said to ‘collapse’, suddenly becoming localised where the particle was detected.

Comparing AWS Lambda performance of Node.js, Python, Java, C# and Go