Lifeboat Foundation

Generative models have had remarkable success in various applications, from image and video generation to composing music and to language modeling. The problem is that we are lacking in theory, when it comes to the capabilities and limitations of generative models; understandably, this gap can seriously affect how we develop and use them down the line.

One of the main challenges has been the ability to effectively pick samples from complicated data patterns, especially given the limitations of traditional methods when dealing with the kind of high-dimensional and complex data commonly encountered in modern AI applications.

Now, a team of scientists led by Florent Krzakala and Lenka Zdeborová at EPFL has investigated the efficiency of modern neural network-based generative models. The study, published in PNAS, compares these contemporary methods against traditional sampling techniques, focusing on a specific class of probability distributions related to spin glasses and statistical inference problems.

The detailed calculations demonstrate that black holes of 10 solar masses may comprise at most 1.2% of dark matter, 100 solar mass black holes—3.0% of dark matter, and 1,000 solar mass black holes—11% of dark matter.

“Our observations indicate that primordial black holes cannot comprise a significant fraction of the dark matter, and simultaneously, explain the observed black hole merger rates measured by LIGO and Virgo,” says Prof. Udalski.

Therefore, other explanations are needed for massive black holes detected by LIGO and Virgo. According to one hypothesis, they formed as a product of the evolution of massive, low-metallicity stars. Another possibility involves mergers of less massive objects in dense stellar environments, such as globular clusters.

On June 25, China’s Chang’e-6 (CE-6) lunar probe is set to return to Earth, carrying the first surface samples collected from the farside of the moon. In anticipation of this historic event, scientists from the Institute of Geology and Geophysics at the Chinese Academy of Sciences are publishing their predictions for the unique materials that may be found in the CE-6 samples in the journal The Innovation.

The space environment is harsh and full of extreme radiation. Scientists designing spacecraft and satellites need materials that can withstand these conditions.

Many people are familiar with oncogenes—genes long known to be involved in cancers in humans, such as the gene Src. What’s less widely understood is that oncogenes didn’t evolve just to cause cancer in species, but rather to control events of normal growth and differentiation.

University of Missouri researchers have developed a way to create complex devices with multiple materials—including plastics, metals and semiconductors—all with a single machine.

The research, which was recently published in Nature Communications, outlines a novel 3D printing and laser process to manufacture multi-material, multi-layered sensors, circuit boards and even textiles with electronic components.

It’s called the Freeform Multi-material Assembly Process, and it promises to revolutionize the fabrication of new products.

Perfume making dates back at least 3,000 years—to the time of Tapputi-belat-ekalle, who is considered the first chemist in history. What we know about her comes from inscriptions on fragments of clay tablets dating back to the Middle Assyrian period (1400–1000BC).

Terahertz technology could help us meet the ever-increasing demand for faster data transfer rates. However, the down-conversion of a terahertz signal to arbitrary lower frequencies is difficult.

In a paper published today in Nature Communications, researchers unveiled previously unobserved phenomena in an ultra-clean sample of the correlated metal SrVO₃. The study offers experimental insights that challenge the prevailing theoretical models of these unusual metals.

The international research team—from the Paul Drude Institute of Solid State Electronics (PDI), Germany; Oak Ridge National Laboratory (ORNL); Pennsylvania State University; University of Pittsburgh; the Pittsburgh Quantum Institute; and University of Minnesota—believes their findings will prompt a re-evaluation of current theories on electron correlation effects, shedding light on the origins of valuable phenomena in these systems, including magnetic properties, high-temperature superconductivity, and the unique characteristics of highly unusual transparent metals.

The perovskite oxide material SrVO₃ is classified as a Fermi liquid—a state describing a system of interacting electrons in a metal at sufficiently low temperatures.