Science News May 10 (Patreon)

[This is a transcript with references.]

Welcome everyone to this week’s science news. Today we’ll talk about Schrödinger’s cat, mind reading with artificial intelligence, a new theory for the origin of life, how many arms the Milky Way has, a design for a new space station, a 3D printed building, 3D ink that conducts electricity, a security system that uses light flashes, and of course, the telephone will ring.

Swiss researchers have made the fattest Schrödinger cat ever.  At least that’s how the headlines put it. They’re alluding to Erwin Schrödinger’s famous thought experiment from 1935 about a cat that’ both dead and alive at the same time. Except it’s a crystal, not a cat. And it’s in a superposition of two oscillation states, not dead and alive. But, well, at least they don’t need to worry about the litter box.

With his thought experiment, Schrödinger wanted to illustrate that quantum mechanics has bigger implications about reality than physicists back then acknowledged. They knew that quantum mechanics allows particles to be in two states at the same time—say left or right—which is called a superposition. They also know that these superpositions only exist until you make a measurement, and then they suddenly “collapse” to one definite outcome. But they thought that’s only the case for small particles. 

Schrödinger argued that if quantum mechanics is right, these superpositions don’t just exist for small things. You could amplify them so that they also exist for large things. Like, for example, cats. Until you measure them.

In Schrödinger’s experiment, a cat is in a box, together with a vial of poison, a trigger mechanism, and a radioactive atom. The nucleus of the atom has a fifty percent chance of decaying in a certain amount of time. If it decays, the trigger breaks the vial of poison, which kills the cat.

But the decay follows the laws of quantum physics. Before you measure it, the nucleus is both decayed and not decayed, and so, it seems that before one opens the box, the cat is both dead and alive. Schrödinger by the way, had a dog.

As you have undoubtedly noticed cats are usually either dead or alive and not both. But just why that is so, physicists still only partly understand. We do understand that tiny interactions with air molecules and even the cosmic microwave background radiation make it impossible to observe superpositions. But they don’t make them go away. Yet, somehow, a measurement does it. That we don’t know how this happens is what’s called the “measurement problem” of quantum mechanics.

That’s why physicists are trying to bring heavier and heavier objects into superpositions so see what happens. For this new paper they used the heaviest object so far. They did it with an oscillating crystal in a cavity coupled to a superconducting circuit. The circuit is a qubit and can be in two different states, depending on how the current flows. It also can be in a superposition of those two states. The researchers then showed that they could transfer the superposition of states from the qubit to the crystal, which was then in two types of oscillations at the same time.

The oscillating crystal is big – 16 micrograms. That’s about 100 quadrillion atoms. Now you might say 16 micrograms isn’t much for a cat and I’d agree, but it’s huge compared to the mass of an electron which is about 10 to the minus 21 micrograms.

It’s also a lot bigger than earlier quantum “cat” experiments. A previous one for example brought 100 million photons into a superposition, so this one’s much of an improvement. But yes, this is physics, where cows are spheres and cats are crystals.

Two papers have got a lot of attention this week, both are about new AI algorithms that can mind read.

The first is from a group of neuroscientists based in Switzerland. They created a deep learning algorithm that can decode what a mouse is seeing and reconstruct a video. For this, they first showed a black and white movie to mice, and no, it wasn’t about cheese. Then collected the neuronal signals from the mice’s visual cortexes as they watched. Then they fed that data into an algorithm and told it what images those signals belonged to, so it could learn to recognize the relation.

Then they showed the movie to a different mouse. The algorithm was then able to translate what the mouse was seeing back into video, frame by frame, at more than 95 per cent accuracy. I know that this looks really impressive, but keep in mind that the algorithm was trained on exactly the same movie. So they did not fully reconstruct what the mouse saw.

They called their algorithm CEBRA, and, yes it’s another one of those smart acronyms. It stands for Consistent EmBeddings of high-dimensional Recordings using Auxiliary variables.  This is the flowchart for how it compresses time series of data to reveal hidden structures within them. It’s really impressive because it’s able to interpret huge amounts of data and can use data from other parts of the body, too.

The second paper is from researchers at the University of Texas at Austin.

They have developed a language decoder that can translate brain activity into text. The study participants first listened to podcasts in an MRI scanner. That way they collected 16 hours of data for each participant and trained the software on that. After that, the person went back into the scanner and silently told a story to themselves, or listened to one. And the decoder translated the brain activity into text.

The algorithm could capture the meaning and often the exact words of the story, as you can see in this chart. The blue bits are exact matches, the purple convey a similar meaning and the red ones are errors.

This graph shows that the accuracy is much better than could be expected by chance. The decoder could also accurately write out descriptions of what the subjects were watching on short videos.

Previous decoders had to be surgically implanted in the brain and they relied on brain signals that relate to movement of the mouth and tongue. For this one you no longer have to imagine you’re voicing the words, but you still have to lie in an MRI scanner, so not like someone’s going to read your mind as you walk down the street. At least for now.

A group of Japanese researchers has come up with a new theory for how life started on Earth.

Scientists have a bunch of theories about how life started on earth. One of the most popular ones has been that lightning sparked the production of molecules that form the basis of DNA and amino acids which are the building blocks of proteins. But this idea requires an atmosphere that contained a lot of methane. But recent work seems to say that the atmosphere on early Earth contained very little methane . Lightning doesn’t have enough energy to make the organic molecules from that. So what was the origin of life?

These researchers now looked, not just at earth, but also at the sun. Observations from other suns like ours tells us that when our sun was young, just about 100 million years old or so, it was about 30 percent fainter than it is today, and it rotated more quickly than it does now. Like most of us at the young age of 100 million, it also had far more frequent and violent outbursts. They could emit up to one thousand times more energy than typical solar flares today.

Those superflares would have spewed out a lot of highly energetic particles, likely every few days, and many of them would have hit young Earth.

So the researchers set about trying to recreate that scenario in the lab. They made an atmosphere of carbon dioxide, molecular nitrogen, water and varying amounts of methane, and then shot at it with protons from Japan’s Tandem particle accelerator.

The found that as long as the methane concentration in the atmosphere was 0 point 5 percent or more, amino acids were formed from their imitation of solar flares. But when they simulated lightning strikes, amino acids were formed only after methane reached 15 percent, much higher than it probably was. We have an entire video about the origin of life coming up on Saturday, so if you’re interested in the topic, mark it on your calendar.

A team of Chinese astrophysicists says that the Milky Way has only two main arms in its spiral.

Until now, scientists thought our galaxy is basically a disk with a blob in the middle and four arms like in this video. But these researchers built up a new picture by crunching data from a whole range of very long baseline interferometry instruments around the world. These experiments use microwaves to measure how far stars are away from Earth.

They also used data from the European Space Agency’s Gaia spacecraft. The Gaia database gave them information on more than 23 thousand massive OB stars. OB stars don’t live for long, so you know they must have formed pretty much where you see them, which is very useful to pin down where star forming regions are. They also looked at nearly one thousand young open star clusters.

Then they plotted them. Here’s a view from the north Galactic Pole, marked with the red circle. The triangles are the stars seen by microwaves. The ‘Perseus arm is in pink. The Norma arm is in orange.  And, no, I didn’t know they had names either.

The researchers now say that these two arms are symmetrical and the dominant structures in the spirals, whereas the other arms are smaller and irregular and probably just messy splits of the two main arms. This makes sense because most spiral galaxies have two arms and with four arms our Milky Way has been somewhat of an outlier. So it turns out our galaxy is quite normal after all. Who knew science news could be so disarming?

The European company Airbus has unveiled plans for a new space station design for NASA.

In January, Airbus joined a public-private partnership to develop the first free-flying commercial space station, called Starlab. These are the first detailed plans. It’s big enough for a crew of four, with three decks, an exercise bay, an air lock and even a centrifuge where astronauts can re-enter gravity to work out. It has greenhouses. And if needed, it could also fly around the moon or to Mars.

Starlab is supposed to be in low-Earth orbit by 2028. It’s meant to be the replacement for the International Space Station, which will be retired in 2031 and crash to Earth. Here are some images from the station this week. That sounds like a pretty good retirement plan to me.

The lead on the Starlab project is the Texas-based company Nanoracks. They got a contract from NASA in 2021, worth 160 million dollar, to build the vessel. NASA is to be its lead tenant, but other nations’ space agencies may also use it as a base. Starlab is one of four spacecraft enterprises NASA has financed. The idea is for the U.S. to dominate the growing commercial use of low-Earth orbit.

It’s in competition with China, which has the space station Tiangong in low-Earth orbit, and India, which plans to launch its space station in 2035. It’s about to get really crowded up there.

Hi Rishi,

Yes, I watched part of the coronation.

Well, it’s important to honour the past. Traditions like this serve to remind people that it still matters what family you were born into.

Ah, nothing says “legitimate transfer of power” quite like placing a big clunky hat on someone’s head.

Talk soon, bye.

Construction has begun on the biggest 3D printed building in Europe. It’s a new data centre in Heidelberg that’s being printed from 450 tonnes of concrete specially made to be printable. This concrete is easy to pump and can be extruded. The finished building will be about 54 meters long, eleven meters wide and nine meters high. It’s supposed to look like this.

Concrete is one of the world’s most common building materials, but it has a huge carbon footprint. It is made by adding gravel and sand to cement. But cement production releases from 560 to 622 kilos of carbon dioxide per tonne. That adds up to about 7 percent of the world’s emissions from energy and industry.

The manufacturer of the new building, Heidelberg Materials, was able to bring the concrete’s carbon footprint down by 55 percent. Plus, the material is 100 percent recyclable, so if it doesn’t work out with the data centre, they can tear it down and instead print a few concrete trees, so we don’t forget what they look like.

An undergraduate at Virginia Commonwealth University has developed printable ink that’s also a good conductor, and no, he didn’t teach squids to wave around a stick and frown at the bassist. It’s a a 3D printable mix of silicone and carbon nanotubes that conducts both heat and electricity, and it sticks to curved surfaces, including skin.

Material scientists have been looking for just such an ink, and indeed developed a few. But the ones they’ve come up with so far have complicated recipes and are cumbersome to use. You have to mix the ingredients together just before you print, and you either need aggressive solvents or have to irradiate the material for a long time to cure it.

The new ink is a one-step material that can be printed at room temperature on glass or latex or even Scotch tape. It uses butyl acetate as a solvent, and it cures in an hour. It’s stretchy and strong, and it conducts electricity.

Here, they attached the ink onto a balloon that simulates a heart, and linked it to electrodes. It’s just a pump, Rohin! The idea is that the ink monitors cardiac function. It distinguishes between normal and abnormal rhythms with changes in the current. At least that’s what they say in the paper.

I’m not much of a doctor but. I was under the impression that chest expansion comes primarily from breathing. And if one can 3D print directly onto your heart, arrhythmia may not be your biggest problem. But ok, leaving aside some anatomical questions, this seems like a really useful invention.

If you had plans to rob a gas station, you probably won’t like this. A neurocognitive scientist at the University of Miami has come up with a security device that scares off attackers by temporarily blinding them.

Say you’re working in a convenience store in the U.S., minding your own business, when a guy with a gun screeches up and tries to rob you. Or, as it’s known in America, Tuesday. You press a button under the counter that flashes thousands of lumens of light, blinding the guy for 10 to 20 seconds. He panics, drops the gun and makes a run for it. In the meantime, the police are on their way.

This security system is called “Lightguard.” Think lifeguard. It’s already got a patent and has been tested by local police departments. There’ve been more than 180 mass shootings in the US this year already, so I welcome this, eh, flash of insight.