Infinite in All Directions is The Wire‘s science newsletter. Click here to subscribe and receive a digest of the most interesting science news and analysis from around the web every Monday, 10 am.
This is a picture of the GSLV F05 mission taking off from the Satish Dhawan Space Centre, Sriharikota, on September 5. If you look closely below the four liquid boosters strapped to the first stage, there are a series of bright spots flanking the main exhaust. I figured these were some kind of vortices until Prateep Basu (who cohosted a live chat on the occasion of the launch for The Wire) corrected me.
They are called shock diamonds – spots of burning gas surrounded by non-burning material. The phenomenon manifests when unburnt gas exiting the engine accumulates in spots of higher pressure, becoming compressed and reignited by the high temperature in these regions.
Shock diamonds form under specific conditions:
- When the engine exhaust is supersonic
- When the pressure of the exhaust gases is lower than the ambient air pressure
The regions of higher pressure form when two things interact: an oblique shockwave caused by the supersonic exhaust and an expansion of the exhaust itself as it travels through the nozzle.
Quantum entanglement is a quirky phenomenon. It predicts that when two particles that have been entangled are separated by a large distance, and a measurement made on one particle, the other particle will change in a fixed way. However, physicists have little clue why this works – even if it always works when tested.
That’s also the case of deep neural networks, a subset of machine learning that uses multiple algorithms to solve problems. The ‘deep’ in the name is derived from the fact that the algorithms are understood to be stacked in many layers, with the input of each layer being the output of the one above it. Another term associated with this architecture is hierarchical feature extraction.
New research from two physicists at Harvard University and the Massachusetts Institute of Technology now claims to be able to work out why their ontology is so meaningful and why neural networks are so efficient at solving problems. The answer seems to lie with how neural networks are similar to the problems they’re good at by being two things: hierarchical and derivative.
From the MIT Tech Review:
… the universe is governed by a tiny subset of all possible functions. In other words, when the laws of physics are written down mathematically, they can all be described by functions that have a remarkable set of simple properties.
So deep neural networks don’t have to approximate any possible mathematical function, only a tiny subset of them.
To put this in perspective, consider the order of a polynomial function, which is the size of its highest exponent. So a quadratic equation like y=x^2 has order 2, the equation y=x^24 has order 24, and so on.
Obviously, the number of orders is infinite and yet only a tiny subset of polynomials appear in the laws of physics. “For reasons that are still not fully understood, our universe can be accurately described by polynomial Hamiltonians of low order,” say Lin and Tegmark. Typically, the polynomials that describe laws of physics have orders ranging from 2 to 4.
The first horseman of Gorakhpur
Every year during the monsoons, hospitals around Gorakhpur, Uttar Pradesh, are flooded with children suffering from acute encephalitis, an inflammatory brain disease, which can be caused by several bacteria, viruses and fungi. However, the actual cause of each annual outbreak has never been properly identified, leaving doctors and hospitals in Gorakhpur at close to breaking point during the rainy season.
Priyanka Pulla writes in Fountain Ink that the region’s administrative officials are convinced the causes are enteroviruses that “spread through nasal, oral and faecal routes”, and that all their efforts to fight the outbreaks have been geared towards eliminating enteroviruses. Whether or not they’re right has been debatable. At the same time, just the fact that the outbreaks persist should mean they’re wrong, right?
Nope. As Pulla explains:
There are two explanations for this failure: That AES isn’t caused by an enterovirus. Or the precautions against enteroviral infections were not properly implemented. Several doctors in Gorakhpur subscribe to the second explanation. Dr. R. N. Singh, who taught at Baba Raghav Das Medical College in the Seventies and began running a campaign for encephalitis eradication after he retired, says the government installed a minuscule 3,000 hand pumps across Gorakhpur, while more than double the number would have been needed to minimise water contamination. “Yeh oonth ke mooh mein jeera bhi nahin hai,” he says, using a Hindi idiom for something so small, that it doesn’t even compare with a cumin seed in a camel’s mouth.
The other explanation for why anti-enterovirus measures aren’t working is that the Gorakhpur illness is caused by scrub typhus, a disease endemic to Uttar Pradesh. Scrub typhus is caused by a bacterium transmitted by a microscopic mite that lives in scrub vegetation. Children walking in bushes without proper clothing are especially at risk, as are woodmen or forest foragers, both common professions in the region.
The evidence for this hypothesis—starting with an August 2014 study by the Indian Council of Medical Research (ICMR) that found scrub typhus antibodies in over 60 per cent of the sick children in Baba Raghav Das—is rapidly growing. But doctors at Baba Raghav Das are sceptical. One reason is that scrub typhus is not known to cause encephalitis outbreaks, triggering a long-drawn fever instead. Encephalitis as a complication occurs in a few untreated sufferers of scrub typhus, rather than among hundreds, as is the case in Gorakhpur, says Komal Kushwaha, a paediatrician who worked at Baba Raghav Das for almost 30 years before retiring in 2015.
Somewhat like Earth
What makes Earth Earth apart from the life on it? In the context of the hunt for exoplanets, the use of the term ‘Earth-like’ usually signals that the planet might be habitable (with huge stress on the ‘might’). However, habitability is the perfect confluence of many, many parameters. And using ‘Earth-like’ to describe a planet even before humankind has developed many of the tools necessary to make detailed observations of it is just misguided.
Mark Kaufman, writing on the Many Worlds site, puts the prematureness of the ‘Earth-like’ tag in context with the things we don’t know about the closest might-be-habitable exoplanet we know: Proxima b, which orbits Proxima Centauri some 4.2 lightyears away.
* The planet, which has a minimum mass of 1.3 Earths and a maximum of many Earths, orbits a red dwarf star. These are the most common class of star in the galaxy, and they put out considerably less luminosity than a star like our sun — about one-tenth of one percent of the power.
* These less powerful red dwarf stars often have planets orbiting much closer to them than what’s found in solar systems like our own. Proxima b, for instance, circles the star in 11.3 days.
* A consequence of this proximity is that the planet is most likely tidally locked by the gravitational forces of the star — meaning that the planet does not rotate like Earth does but rather has a daytime and nighttime side like our moon. Some now argue that a tidally locked planet could theoretically be habitable, but the consensus seems to be that it is an obstacle to habitability rather than a benefit.
* The authors of the Proxima b paper make clear that evidence that the planet is rocky (as opposed to gaseous) is limited, and that’s why they label it as a “candidate terrestrial planet.”
These uncertainties are interesting because, if we don’t find a way to ascertain their truth-value, Proxima b becomes one of various exoplanets that possess many of its other features but are evidently uninhabitable. As Kaufman writes,
So to describe Proxima b as “Earth-like” seems unfortunate to me, and prone to giving the public the misguided impression of a planet with blue skies, oceans, and fish swimming in them. Proxima b may have some very broadly defined characteristics that parallel Earth, but so do many other exoplanets that are definitely not habitable. And therein lies the really interesting part.
The operational word in the details surrounding North Korea’s fifth nuclear test, according to Jeffrey Lewis writing in Foreign Policy, is “standardised”. Contingent to the veracity of the claim, made by Kim Jong Un, Lewis says that if DPRK has indeed developed standardised nuclear warheads and missiles that can carry them, they can effectively now mass produce the things.
And that’s where the problem begins. Lewis brings together various seemingly disparate claims made over the years by Kim to yield a highly explosive geopolitical charge set off by two things (in his words):
[The North Koreans] use composite pits of both Pu and [highly enriched uranium] (mixed charge) and they ‘boost’ the yield of the explosion using a gas of hydrogen isotopes (prompt thermonuclear reaction).
Based on various estimates of how much plutonium and uranium the state possesses, that means at least 20 nuclear warheads in DPRK’s possession. And these aren’t puny either: the yield from the fifth test was equivalent to the bomb dropped on Nagasaki on August 9, 1945.
Slow… but steady?
The 30th magnetar to be discovered also seems to be the slowest spinning one ever found. A magnetar is a neutron star with an intense magnetic field enveloping it. All neutron stars, including magnetars, found to date have been known to spin with a period around a few seconds. They are the remnants of heavy stars that went supernova and then imploded, forcing their cores to become compressed into extremely small spaces. The resulting neutron star is sent spinning because that’s a way to lose energy and become more stable.
But oddly, 1E 1613 seems to have a period of 24,000 seconds (6.6 hours). It is located 9,000 lightyears from Earth, and is the remnant of a star that went kablooey 2,000 years ago. Astronomers think that two millennia is hardly enough time for a magnetar to go from spinning once every 10 seconds to once every 24,000. Theories are emerging to explain the anomaly – one being that magnetic material expelled from the originator star fell back into the magnetar, slowing it down.
NASA has more details.
Like what you read? Subscribe to the newsletter.