BBlogFlow

Surprisingly high redundancy in electronic structure data across materials explained by low intrinsic dimensionality - Nature

·1h ago

This story deserves to be shared

Send it to someone who'd care, or post it on your networks.

Facebook Twitter Email

AcknowledgementsThe authors thank UCLA's Institute for Digital Research and Education (IDRE), Michigan Technological University's MRI computing clusters, and the National Energy Research Scientific Computing Center (NERSC) for providing computational resources. The authors thank Abhishek Keripale for assistance with the development of later versions of the CrFeCoNi machine-learning models. The authors used GPT-5 (OpenAI) for language editing and proofreading and subsequently reviewed and edited all content. FundingThis work was primarily supported by the U. S. Department of Energy (DOE), Office of Science, Basic Energy Sciences (BES), under Award DE-SC0023432, which provided primary support for the overall project conception and integrated KS-DFT–ML investigations. The University of California National Laboratory Fees Research Program under Grant L25CR9003 supported data generation and post-processing at UCLA (ASB). The National Science Foundation under Award 2442313 supported data-pruning and dimensionality analyses at Michigan Technological University (SG, SH). This research also used resources of the National Energy Research Scientific Computing Center (NERSC), a U. S.

This article is an excerpt. To read the full content, visit the original source. Nature

Reactions

0 comments

No comments yet. Be the first to share your thoughts!

Facebook Twitter

Related articles

Patient-Owned Health Data Model Reduces Clinical Trial Recruitment Redundancy - The Clinical Trial Vanguard

The same patient, identified and re-consented for the third or fourth time across separate studies, is not an edge case. It is how clinical research is currently structured, and the inefficiency is measurable: every redundant records request, re-registration, and repeat consent conversation adds cost to the sponsor and time to a patient already managing a demanding diagnosis. The case building around patient-owned health data argues that the architecture itself is the problem, not the execution. The traditional model treats each protocol as a discrete relationship. A trial opens, participants are found, data flows inward, and the relationship ends. But sponsors rarely stop needing evidence from the same populations once a study closes. Observational work, real-world outcomes, label extensions, and post-market commitments all draw on the same patient groups, rebuilt from scratch each time. What the patient-owned model proposes instead is a single consented relationship that persists across those activities, turning what was an episodic research interaction into ongoing infrastructure.

The Clinical Trial Vanguard·

Some Planetary Systems Are Born for Chaos. Astronomers Think They Know Why - ZME Science

Illustration of the genetic link between stellar properties and planetary system architectures. Credit: INAF / Universitat de València / image generation. For most of human history, planetary systems were a sample size of one. Ours. Then people learned about Mercury, Venus, Mars, Jupiter, and Saturn since antiquity, tracking their movements for generations. Uranus joined the planetary roster in 1781 thanks to the telescope, followed by Neptune in 1846, but every planet we discovered still belonged to the same family orbiting our Sun.

ZME Science·

NASA confirms it: the Moon is now smaller - MARCA

08:02CDTA study funded by NASA, and published in The Planetary Science Journal, has determined that the Moon's radius has decreased by about 45 meters. The scientists' analysis concluded that this reduction occurred over the last 400 to 800 million years. Every million years, according to the report's estimates, the Earth’s satellite’s radius would shrink by between 6 and 11 centimeters. In other words, 0.1 micrometers (0.0001 millimeters) per year. Aside from this fact, the research funded by the U. S. space agency has identified more than 10,000 faults on the Moon as a result of studying images from the Lunar Reconnaissance Orbiter, in orbit since 2009. This finding could pose a setback for exploration of the satellite: NASA is considering establishing a lunar base with basic infrastructure in 2032. Until then, efforts are focused on determining possible locations for the installation. The large number of faults detected goes beyond mere scientific curiosity, as it could pose a setback for future crewed missions. The reason is that, unlike the Earth, with 15 major tectonic plates, the Moon has a single plate.

MARCA·