Heating up Twitter feeds this week was a new commentary about a rarely discussed perspective on autism, talk about the use of optogenetics tools to manipulate ‘presynapses,’ a possible explanation for decision-making differences between the sexes, and a striking illustration of the human brain.
In most sci-fi settings, the writer doesn’t have magic to work with, so unstoppable machines too small to see are often the next best thing.
😳!!!
The Moon isn’t necessarily there if you don’t look at it. So says quantum mechanics, which states that what exists depends on what you measure. Proving reality is like that usually involves the comparison of arcane probabilities, but physicists in China have made the point in a clearer way. They performed a matching game in which two players leverage quantum effects to win every time—which they can’t if measurements merely reveal reality as it already exists.
“To my knowledge this is the simplest [scenario] in which this happens,” says Adan Cabello, a theoretical physicist at the University of Seville who spelled out the game in 2001. Such quantum pseudotelepathy depends on correlations among particles that only exist in the quantum realm, says Anne Broadbent, a quantum information scientist at the University of Ottawa. “We’re observing something that has no classical equivalent.”
Continue reading “Reality doesn’t exist until you measure it, quantum parlor trick confirms” »
At the time climate change was only beginning to be talked about in the scientific community as well as behind the scenes among researchers working for fossil fuel companies.
Climate change fit the EPA’s mandate. And unlike an oil or chemical spill, no reputable scientist would see climate change as equivalent to “the crisis of the day.” But this phrase appears in Chief Justice John Roberts’ opinion justifying the decision in West Virginia v. EPA to deny the Agency its power to regulate carbon emissions from coal-fired power plants which based on the mandated powers described above is its purview (see points 3, 4, and 5).
The design of protein sequences that can precisely fold into pre-specified 3D structures is a challenging task. A recently proposed deep-learning algorithm improves such designs when compared with traditional, physics-based protein design approaches.
ABACUS-R is trained on the task of predicting the AA at a given residue, using information about that residue’s backbone structure, and the backbone and AA of neighboring residues in space. To do this, ABACUS-R uses the Transformer neural network architecture6, which offers flexibility in representing and integrating information between different residues. Although these aspects are similar to a previous network2, ABACUS-R adds auxiliary training tasks, such as predicting secondary structures, solvent exposure and sidechain torsion angles. These outputs aren’t needed during design but help with training and increase sequence recovery by about 6%. To design a protein sequence, ABACUS-R uses an iterative ‘denoising’ process (Fig.
Instead of predicting the shapes of naturally occurring molecules, software designs original ones.
