Lifeboat Foundation

The idiom “actions speak louder than words” first appeared in print almost 300 years ago. A new study echoes this view, arguing that combining self-supervised and offline reinforcement learning (RL) could lead to a new class of algorithms that understand the world through actions and enable scalable representation learning.

Machine learning (ML) systems have achieved outstanding performance in domains ranging from computer vision to speech recognition and natural language processing, yet still struggle to match the flexibility and generality of human reasoning. This has led ML researchers to search for the “missing ingredient” that might boost these systems’ ability to understand, reason and generalize.

In the paper Understanding the World Through Action, UC Berkeley assistant professor in the department of electrical engineering and computer sciences Sergey Levine suggests that a general, principled, and powerful framework for utilizing unlabelled data could be derived from RL to enable ML systems leveraging large datasets to better understand the real world.

A huge asteroid the size of the Eiffel Tower is approaching our planet. The asteroid is named 4,660 Nereus and has been flagged ‘Potentially Hazardous’ by NASA. Nereus is 330 meters long and will break into Earth’s orbit on Saturday, December 11. The colossal asteroid is traveling at 23,700 km/h towards our planet.

On December 11, the asteroid will make its closest approach to Earth. It will come within 3.86 million km, about ten times the distance between Earth and the Moon. Although it sounds like an enormous gap on cosmic scales, it is actually a stone’s throw away.

Continue reading “A Stadium Sized Asteroid Is Approaching Us This Weekend | Everything You Need To Know | Nereus” »

To set some benchmarks for their simulator, the researchers tried out three different design algorithms working in conjunction with a deep reinforcement learning algorithm that learned to control the robots through many rounds of trial and error.

The co-designed bots performed well on the simpler tasks, like walking or carrying things, but struggled with tougher challenges, like catching and lifting, suggesting there’s plenty of scope for advances in co-design algorithms. Nonetheless, the AI-designed bots outperformed ones design by humans on almost every task.

Continue reading “Robots Evolve Bodies and Brains Like Animals in MIT’s New AI Training Simulator” »

China has succeeded in literally changing the weather, a new study claims.

As the South China Morning Post reports, researchers say that during the Chinese Communist Party’s centennial celebration over the summer, weather authorities successfully modified the weather above Beijing to clear the sky and reduce pollution for the tens of thousands gathered for a commemorative ceremony in Tiananmen Square.

They did so using cloud seeding technology, a long-studied but controversial process that involves shooting silver iodide particles into the clouds, with the idea of attracting water droplets to modify the weather.

Neoen reveals plans for another big battery in South Australia which could be even bigger than the Victorian Big Battery opened on Wednesday.

French renewable energy and battery storage developer Neoen, fresh from the formal opening of the Victorian Big Battery – the biggest in Australia to date – has revealed plans for a potentially even bigger battery in South Australia.

Neoen is due to hold a community open day on Thursday for the Blyth battery, which will be sized up to 300MW and 800MWh, trumping the 300MW/450MWh capacity of the newly opened VBB near Geelong.

Continue reading “Neoen reveals plans for another 300MW big battery in South Australia” »

Designed for high resolution from load cells and high amperage on the MOSFETs, this automation board packs a punch.

With the Zappi, Myenergi offers a compact charging station that can charge the electric vehicle in the garage or carport at home with solar power from the roof without any additional components.

Researchers from the Skolkovo Institute of Science and Technology and Saratov State University have come up with an inexpensive method for visualizing blood flow in the brain. The new technique is so precise it discerns the motions of individual red blood cells — all without the use of toxic dyeing agents or expensive genetic engineering. The study was published in The European Physical Journal Plus.

To understand more about how the brain’s blood supply works, researchers map its blood vessel networks. The resulting visualizations can rely on a variety of methods. One highly precise technique involves injecting fluorescent dyes into the blood flow and detecting the infrared light they emit. The problem with dyes is they are toxic and also may distort mapping results by affecting the vessels. Alternatively, researchers employ genetically modified animals, whose interior lining of blood vessels is engineered to give off light with no foreign substances involved. Both methods are very expensive, though.

Researchers from Skoltech and Saratov State University have devised an inexpensive method for visualizing even the smallest capillaries in the brain. The method — which integrates optical microscopy and image processing — is dye-free and very fine-grained, owing to its ability to detect each and every red blood cell travelling along a blood vessel. Since the number of RBCs in capillaries is not that high, every cell counts, so this is an important advantage over other methods, including dye-free ones.

None.

Researchers at Human Brain Project partner University of Granada in Spain have designed a new artificial neural network that mimics the structure of the cerebellum, one of the evolutionarily older parts of the brain, which plays an important role in motor coordination. When linked to a robotic arm, their system learned to perform precise movements and interact with humans in different circumstances, surpassing performance of previous AI-based robotic steering systems. The results have been published in the journal Science Robotics.