Lifeboat Foundation: Safeguarding Humanity
Toggle light / dark theme

Blog

Category: information science

A trio of AI researchers at Google’s Google DeepMind, working with a colleague from the University of Toronto, report that the AI algorithm Dreamer can learn to self-improve by mastering Minecraft in a short amount of time. In their study published in the journal Nature, Danijar Hafner, Jurgis Pasukonis, Timothy Lillicrap and Jimmy Ba programmed the AI app to play Minecraft without being trained and to achieve an expert level in just nine days.

Over the past several years, computer scientists have learned a lot about how can be used to train AI applications to conduct seemingly intelligent activities such as answering questions. Researchers have also found that AI apps can be trained to play games and perform better than humans. That research has extended into , which may seem to be redundant, because what could you get from a computer playing another computer?

In this new study, the researchers found that it can produce advances such as helping an AI app learn to improve its abilities over a short period of time, which could give robots the tools they need to perform well in the real world.

Read more | >

A team of physicists have discovered a new approach that redefines the conception of a black hole by mapping out their detailed structure, as shown in a research study recently published in Journal of High Energy Physics.

The study details new theoretical structures called “supermazes” that offer a more universal picture of to the field of theoretical physics. Based in , supermazes are pivotal to understanding the structure of black holes on a microscopic level.

“General relativity is a powerful theory for describing the large-scale structure of black holes, but it is a very, very blunt instrument for describing black-hole microstructure,” said Nicholas Warner, co-author of the study and professor of physics, astronomy and mathematics at the USC Dornsife College of Letters, Arts and Sciences. In a framework of theories extending beyond Einstein’s equations, supermazes provide a detailed portrait of the microscopic structure of brane black holes.

Read more | >

Monash University researchers have extended Descartes’ Circle Theorem by finding a general equation for any number of tangent circles, using advanced mathematical tools inspired by physics. A centuries-old geometric puzzle dating back to the 17th century has finally been solved by mathematicians

Read more | >

Big data has gotten too big. Now, a research team with statisticians from Cornell has developed a data representation method inspired by quantum mechanics that handles large data sets more efficiently than traditional methods by simplifying them and filtering out noise.

This method could spur innovation in data-rich but statistically intimidating fields, like and epigenetics, where traditional data methods have thus far proved insufficient.

The paper is published in the journal Scientific Reports.

Read more | >

A new large language model framework teaches LLMs to use an optimization solving algorithm to resolve complex, multistep planning tasks. With the LLMFP framework, someone can input a natural language description of their problem and receive a plan to reach their desired goal.

Read more | >

The interactions between light and nitroaromatic hydrocarbon molecules have important implications for chemical processes in our atmosphere that can lead to smog and pollution. However, changes in molecular geometry due to interactions with light can be very difficult to measure because they occur at sub-Angstrom length scales (less than a tenth of a billionth of a meter) and femtosecond time scales (one millionth of a billionth of a second).

The relativistic ultrafast electron diffraction (UED) instrument at the Linac Coherent Light Source (LCLS) at SLAC National Accelerator Laboratory provides the necessary spatial and time resolution to observe these ultrasmall and ultrafast motions. The LCLS is a Department of Energy (DOE) Office of Science light source user facility.

In this research, scientists used UED to observe the relaxation of photoexcited o–nitrophenol. Then, they used a genetic structure fitting algorithm to extract new information about small changes in the molecular shape from the UED data that were imperceptible in previous studies. Specifically, the experiment resolved the key processes in the relaxation of o-nitrophenol: proton transfer and deplanarization (i.e., a rotation of part of the molecule out of the molecular plane). Ab-initio multiple spawning simulations confirmed the experimental findings. The results provide new insights into proton transfer-mediated relaxation and pave the way for studies of proton transfer in more complex systems.

Read more | >

Researchers at Osaka University have revealed a link between the equations describing strain caused by atomic dislocations in crystalline materials and a well-established formula from electromagnetism, an insight that could advance research in condensed matter physics. A fundamental goal of physi

Read more | >