Lifeboat Foundation

Reservoir computing is a promising computational framework based on recurrent neural networks (RNNs), which essentially maps input data onto a high-dimensional computational space, keeping some parameters of artificial neural networks (ANNs) fixed while updating others. This framework could help to improve the performance of machine learning algorithms, while also reducing the amount of data required to adequately train them.

RNNs essentially leverage recurrent connections between their different processing units to process sequential data and make accurate predictions. While RNNs have been found to perform well on numerous tasks, optimizing their performance by identifying parameters that are most relevant to the task they will be tackling can be challenging and time-consuming.

Jason Kim and Dani S. Bassett, two researchers at University of Pennsylvania, recently introduced an alternative approach to design and program RNN-based reservoir computers, which is inspired by how programming languages work on computer hardware. This approach, published in Nature Machine Intelligence, can identify the appropriate parameters for a given network, programming its computations to optimize its performance on target problems.

GPS is now a mainstay of daily life, helping us with navigation, tracking, mapping, and timing across a broad spectrum of applications. But it does have a few shortcomings, most notably not being able to pass through buildings, rocks, or water. That’s why Japanese researchers have developed an alternative wireless navigation system that relies on cosmic rays, or muons, instead of radio waves, according to a new paper published in the journal iScience. The team has conducted its first successful test, and the system could one day be used by search and rescue teams, for example, to guide robots underwater or to help autonomous vehicles navigate underground.

“Cosmic-ray muons fall equally across the Earth and always travel at the same speed regardless of what matter they traverse, penetrating even kilometers of rock,” said co-author Hiroyuki Tanaka of Muographix at the University of Tokyo in Japan. “Now, by using muons, we have developed a new kind of GPS, which we have called the muometric positioning system (muPS), which works underground, indoors and underwater.”

Summary: Researchers mapped neural activity in an octopus’s visual system, revealing striking similarities to humans.

The team observed neural responses to light and dark spots, thereby creating a map resembling the organization of the human brain. Interestingly, octopuses and humans last shared a common ancestor around 500 million years ago, suggesting independent evolution of such complex visual systems.

These findings contribute greatly to our understanding of cephalopod vision and brain structure.

Lot’s of science news, stay till the end for the climate stuff.

Expand your scientific horizon with Brilliant! Use our link https://brilliant.org/sabine You can get started for free, and the first 200 will get 20% off the annual premium subscription.

Continue reading “Record Temperatures in the North Atlantic” »

A key challenge in neuroscience is to understand how the brain can adapt to a changing world, even with a relatively static anatomy. The way the brain’s areas are structurally and functionally related to each other—its connectivity—is a key component. In order to explain its dynamics and functions, we also need to add another piece to the puzzle: receptors.

Now, a new mapping by Human Brain Project (HBP) researchers from the Forschungszentrum Jülich (Germany) and Heinrich-Heine-University Düsseldorf (Germany), in collaboration with scientists from the University of Bristol (UK), New York University (U.S.), Child Mind Institute (U.S.), and University of Paris Cité (France) had made advances on our understanding of the distribution of receptors across the brain.

The findings were published in Nature Neuroscience, and the data is now freely available to the neuroscientific community via the HBP’s EBRAINS infrastructure.

Superfast, subatomic-sized particles called muons have been used to wirelessly navigate underground for the first time. By using muon-detecting ground stations synchronized with an underground muon-detecting receiver, researchers at the University of Tokyo were able to calculate the receiver’s position in the basement of a six-story building.

As GPS cannot penetrate rock or water, this new technology could be used in future search and rescue efforts, to monitor undersea volcanoes, and guide autonomous vehicles underground and underwater. The findings are published in the journal iScience.

GPS, the global positioning system, is a well-established navigation tool and offers an extensive list of positive applications, from safer air travel to real-time location mapping. However, it has some limitations. GPS signals are weaker at higher latitudes and can be jammed or spoofed (where a counterfeit signal replaces an authentic one). Signals can also be reflected off surfaces like walls, interfered with by trees, and can’t pass through buildings, rock or water.

Early Dark Energy Spectroscopic Instrument (DESI) release holds nearly two million objects, including distant galaxies, quasars and stars in our own Milky Way.

Dark Energy Spectroscopic Instrument (DESI), the most robust multi-object survey spectrograph, capable of mapping more than 40 million galaxies, quasars, and stars, recorded an 80-terabyte data set this Tuesday.

Continue reading “First batch of DESI data available for scientists to mine” »

Photosynthesis. The maps elucidate the complex energy transfer process in photosynthesizing bacteria, providing a clear picture of how sunlight energy is channeled from the outer to the inner molecular ring of the light-harvesting complex.

Systems obeying quantum mechanics are notoriously difficult to visualize, but researchers at the University of Illinois Urbana-Champaign have developed an illustration technique that displays quantum features in an easy-to-read diagram called a coherence map. The researchers used these maps to study the quantum mechanisms that underlay photosynthesis, the process by which plants and some bacteria use sunlight to convert carbon dioxide and water into food.

This image of galaxy cluster MACS J1206.2–0847 (or MACS 1,206 for short) is part of a broad survey with NASA’s Hubble Space Telescope.

The distorted shapes in the cluster are distant galaxies from which the light is bent by the gravitational pull of an invisible material called dark matter within the cluster of galaxies. This cluster is an early target in a survey that will allow astronomers to construct the most detailed dark matter maps of more galaxy clusters than ever before.

These maps are being used to test previous, but surprising, results that suggest that dark matter is more densely packed inside clusters than some models predict. This might mean that galaxy cluster assembly began earlier than commonly thought.