Sabine
patreon
Science News Nov 30
🕑 Added 2022-11-30 16:00:10 +0000 UTC[This is a transcript with references]
Welcome everyone to this week’s science news! Today we’ll talk about Trouble at ITER, robots that build robots, air pollution, AI that classifies supernovae, a small asteroid that hit Canada, Super GPS, a new supercomputer simulation of the sun, a quantum thermometer. And of course, the telephone will ring.
The International Thermonuclear Experimental Reactor, ITER, for short, recently announced that they found major defects in two key components of the machine, likely causing a costly and time-intensive setback.
Already in November 2021, ITER engineers found a leak in an element of the thermal shield of the vacuum vessel that had been delivered in 2020. But the scope of the problem only became apparent now. It turned out that one of the helium pipes that connect to the vessel didn’t fit smugly enough, causing cracks to develop. The vacuum vessel is made of four individual segments but according to the press release “deviations from nominal dimensions were more substantial than the specified limit in different locations on the component’s outer shell”. As a consequence, welding the pieces together caused stress which caused cracks, and with helium leaking through the cracks, that caused corrosion.
ITER is the biggest international nuclear fusion project currently underway. It uses the tokamak design and construction has been underway in Cadarache, France, since 2010. ITER’s goal is to produce net energy in the reactor vessel, though, if one takes into account the entire energy necessary to operate the reactor, ITER will still not produce net power.
The first plasma run in the ITER vessel was originally planned for 2018, but in 2016 the ITER council pushed it back to 2025. The newly discovered defects are likely to further delay operations.
Representatives from the thirty-five countries involved in the ITER project met in a hybrid format on November 16th and 17th and urged for repair to start as soon as possible.
I am very supportive of research on nuclear fusion, but sometimes it’s really hard.
Scientists at MIT have made a big step forward in creating robots that can build nearly anything, including bigger robots, according to a paper that was just published in Nature Communications Engineering.
Researchers had previously developed small units from lightweight materials that can be assembled by robots to anything from airplane wings to racing cars. Those units are called “voxels”, that’s a three-dimensional version of pixels. The MIT researchers have now made these construction voxels more complex, so that they can carry out work themselves. While the current research did not actually see robots build new robots, it shows that the pieces are all working as intended.
This automated assembly didn’t just require mechanical elements, it also required new software solutions because the robots must decide themselves what piece goes where and in which case to build a new robot. That too, seems to be working as desired. At present, the construction units are quite bulky, but this research is really just a proof of principle, and the technology can be miniaturized once it works.
Don’t worry too much about it, those robots still need humans to provide the building blocks. At least for now.
A new study in the journal Nature Medicine has found that air pollution impairs the immune system.
The study, conducted by immunologists at Columbia, wasn’t originally looking for a link between air pollution and problems of the immune system. They began ten years ago to collect tissues from deceased organ donors to study immune cells. But when they found pollutants in the lymph nodes, they shifted their research interests.
They found that the older the donor, the more pollutants they had in the lymph nodes. Those from children and teenagers were largely beige, while those from donors over the age of 30 were increasingly tinged with black. There are photos in the paper, but I’m not showing them because they’re not very appetizing and YouTube is funny with that kind of thing. It’s probably for the better.
More specifically, the researchers found that the pollutants build up inside macrophages, which are immune cells that engulf and destroy bacteria, viruses, cellular debris, and other potentially dangerous substances. According to Donna Farber who led the study: “These immune cells are simply choked with particulates and could not perform essential functions that help defend us against pathogens.” The results add to the growing body of evidence that air pollution is a much bigger health hazard than scientists appreciated only a few decades ago.
Hi Rishi,
Ah I’m sure the people of Great Britain will understand the need to pay taxes for, erm, your garden statue.
I mean, they can’t eat bronze anyway. Right. Bye.
Astronomers at Caltech have developed a machine learning algorithm that has autonomously classified one thousand supernovae. The algorithm worked on data gathered by the Zwicky Transient Facility, ZTF for short, located at Caltech's Palomar Observatory. The supernovae algorithm began its work last year, and within just a year and a half has made it to 1000 supernovae classifications.
ZTF scans the skies every night in search of “transient events”, that everything from asteroids to black hole collisions to supernovae. Each night, the instrument sends out hundreds of thousands of alerts to astronomers around the world. This is way too much data for scientists to track and classify manually and risks that important data might go unnoticed just because there aren’t sufficiently many humans looking at it. The supernovae algorithm greatly alleviates this problem.
In the past decade, artificial intelligence has had a major impact on science, and the trend shows no signs of slowing down. According to a new report from Australia’s National Science Agency, artificial intelligence is now being used in ninety-eight percent of all scientific fields. On average, about 5 point 7 percent of all peer-reviewed research worldwide makes use of AI. The physical sciences are leading the pack with more than 10 percent of all papers now using artificial intelligence.
Contrary to rumours, I am totally real.
A small asteroid hit the sky over south-west Canada on the morning of November 19th. It broke up when it entered the atmosphere and likely scattered some small meteorites over the southern coastline of Lake Ontario.
According to a press release from NASA their impact hazard assessment system, called “Scout” worked as desired. The asteroid was discovered on the evening of November 18 by the Catalina Sky Survey, which is headquartered at the University of Arizona in Tucson. They assigned the object the name “WJ1 2022”.
Just seven minutes after they had reported it to NASA, the Scout system had determined that the asteroid had a 25 percent probability of hitting Earth's atmosphere, with possible impact locations spanning across North America.
NASA then used further tracking data to continually update the asteroid’s trajectory and re-assess the chance of impact. Observations within an hour of first discovery raised the probability of impact to 100 percent. They constructed a 3-d model and predicted the impact area to high precision. In total, NASA collected 46 observations of the asteroid’s position from telescopes all over the planet.
According to Kelly Fast, the program manager of the Near Earth Observations program at NASA “The planetary defense community really demonstrated their skill and readiness with their response to this short-warning event.”
Hi Elon,
No, it’s a good idea to remove the character limit, I think that way, other people will be much easier to ignore.
See you later.
The Global Positioning System, GPS, has become so widely used that by now many of us have no idea how to find directions without it. But sometimes the GPS is unreliable, especially in dense urban areas where it’s difficult to get a straight line of sight to a satellite. Researchers at Delft University in the Netherlands are working on a new positioning system that could determine the position of a receiver in urban areas to within 10 centimeters. They have called it, drums please, SuperGPS.
The new system would not use satellites, but a grid of radio transmitters that are synchronized to extremely high accuracy through an optical telecommunication network.
Jeroen Koelemeij from the SuperGPS team says: “We realized that with a few cutting-edge innovations, the telecommunication network could be transformed into a very accurate alternative positioning system that is independent of GPS.” They just recently performed a test run with six radio transmitters. It all worked as desired, and they published their results in Nature.
The paper also provides further technical details for how the system would work that I’ll not explain because I don’t understand them. I do understand however that with 10 centimetres accuracy even I might finally be able to find the regent’s canal.
NASA performed a new supercomputer simulation for a cycle of sunspot formation. Researchers at the Advanced Supercomputing facility at NASA’s Research Center in Silicon Valley have once again studied the dynamical behaviour of the sun. The sun might look simple as it sits there in the sky, patiently shining down on us, but its plasma has very complex and partly chaotic dynamics whose details aren’t well understood. Besides the solar cycles, whose exact nature and origin haven’t been entirely sorted out, one of the biggest mysteries are “torsional oscillations” of magnetic fields that are involved in the formation of sun spots and solar flares.
A computational analysis of data from two NASA spacecraft now revealed that the strength of those torsional oscillations is currently declining which indicates that the current solar cycle is weaker than the previous one. Changes in the sun’s behaviour impact space weather, the Earth’s atmosphere, and in the long run, also our climate.
NASA has made supercomputing a key element of their endeavours. Whether it be for the design of safer, more efficient air taxi systems, keeping planetary rovers safe during risky landings, helping engineers model heat shiels, predicting weather and climate, or exploring planets inside and outside our solar system, supercomputing helps. Maybe one day they’ll even manage to explain how governments of so-called advanced nations sometimes find budget holes of several billion dollars.
I haven’t talked about any quantum stuff so far which isn’t in the natural order of things, so here’s the best quantum paper of the week according to Sabine. It was just published in PRL and comes from a group of researchers in Australia and Japan. They propose a new way of measuring the Unruh effect.
The Unruh effect, named after Bill Unruh, is a so far hypothetical vacuum temperature that should be seen by an observer who accelerates. It would mean that if you accelerate, then the vacuum is no longer a vacuum, but it contains real particles.
The reason this is at least mathematically possible is that if you combine quantum mechanics with special relativity, then the notion of a particle itself becomes observer-dependent. If you accelerate a detector, it should detect a temperature that depends on the acceleration. I have never been entirely convinced the Unruh effect is physically real, so I’d really like to see it tested. Unfortunately, it’s incredibly difficult to test because the acceleration you’d have to reach is far off experimental possibilities.
The authors of the new paper now say you don’t need high acceleration to see the Unruh effect, you can see the same effect if you use a detector that interacts with a time-dependent electromagnetic field that changes the energy gap of the detector. Now of course that is no longer a vacuum, but this configuration would allow the detector to become sensitive to the vacuum contribution. And they say that the size of the effect is larger than that of the usual Unruh effect, and less sensitive to noise, so it might be possible to measure it in the near future.
I am not entirely convinced that this is actually physically the same as the Unruh effect. But in any case, the authors point out that one could also use this method the other way round, namely to detect small temperatures. I expect this paper to lead to a lot of follow-up work in the near future.
I think this is a really cute paper by which I really mean I wish I’d thought of it first. Even Albert is impressed.
Comments
Fusion power is always five years out, right?
About the SARS-CoV-2 vaccines' safety and efficacy, I suggest that you go to This Week in Virology (at microbe.tv/twiv) to find out that the vaccines are indeed safe and effective. Dr. Daniel Griffen discusses the latest clinical findings, including that there are no substantiated deaths from the vaccine and that the vaccines are safe and effective, continually. Coincidently, in his latest live stream (https://www.youtube.com/watch?v=nCzsWtlogTc), Vincent Racaniello and Amy Rosenfeld discussed a paper that discussed alleged deaths from myocarditis after vaccination (starting at ~7:55). Spoiler alert, the authors admit that they cannot prove a causal link. The evidence supporting the anti-vaxxers' claims that the SARS-CoV-2 vaccine is sorely lacking and the data showing that those vaccinated suffer reduced severe illness and death is overwhelming.
I got your message about covidiots... I would rather take a more charitable approach towards most of them, like Sabine did with the flat-earthers. They are not stupid, just mislead, without the necessary tools to properly analyze data. That's where Sabine steps in to help. I don't have a charitable view as to one of the main covid skeptics, who shall remain unnamed. I think he intentionally misleads, perhaps for financial gain, or other reasons.
@Tracey I've been meaning to reply to you but somehow I'm making comments off-thread.
Try shift-return for a new paragraph. Trial and error got that sorted for me. Good point about social science data vs hard science data. I hadn't considered that at all.
I just deleted my first comment, since it was unnecessary.
Hi Luval, my apologies. We have had at least one Covidiot in Patreon. I need to slow down and actually read comments properly.
I might mail myself zu Deutschland, however. 🤔
I'm in Sydney, there are ibis almost everywhere. The one in my Twitter profile was in Potts Point, near the CBD. I'm not going to England just to be blown up by geese! Meine Katze is an adequate ESP for the time being.
Stop being a dickhead Colleen! Instead, make your way to Regent's canal to search for your next emotional support Ibis. Don't blow up any barges while you're there. Pick up some right, proper English cheese before heading back down under. Don't mail yourself to London or back to Sydney as the trip could take a full month each way :-).
The data in the covid skeptic case are not misrepresented, but misinterpreted. But the misinterpretation becomes obvious once the confounds are pointed out, or once background levels (from natural deaths having nothing to do with vaccine) are reported. Thanks Tracey for tip on new paragraphs. Episode could be for other data besides covid data, but covid skeptics offer us so many juicy examples of bad data interpretation, and this is why many people don't even look at anything but randomized, Placebo-controlled clinical trials. It would be highly educational and also controversial, since Sabine might like controversy? Also, social science data is much more prone to this kind of error, something we did not learn about in our physics education, and some engineers who are covid skeptics also did not...
Hi Luval and Colleen, I think the point is that an analysis looks like it says one thing but because of biases in the analysis, the data are misrepresented. Some of the covid gibberish is given for examples. I think this would be a good topic for an episode as well, but it need not be covid specific, it could just outline many ways in which data are analyzed or presented in a biased way.
I made an appeal against my ban but no joy. Now I'll have to figure the canal thing out myself if Sabine or someone else doesn't enlighten us. Those geese were terrible at emotional support.
I got a SIM card from Aldi the same day I picked up my new phone and used it to make a new account. 😀 I'm not being silly enough to post words describing certain body parts as insults again, however I don't think that's anywhere near as nasty and mean as those who were making tweets composed of racist slurs because they could. You shall know me by the picture of the Ibis next to the fountain pond that is my profile picture still.
Can you do a video on graphs that seem to show that the covid vaccines are unsafe and/or ineffective, until one analyzes background levels (higher than death reports) and confounds (death following covid shot reporting is legally mandatory whereas seasonal flu shot deaths is not)? Or another example, epidemiological data from US CMS that seems to suggest a 2x higher death rate from covid vaccine, until one realizes that the covid vaccine comes in two rounds? And then showing randomized, placebo-controlled trials showing safety and efficacy.
I'm sorry, that information is reserved exclusively for people who have not gotten kicked off of twitter :-). Now that Elon's in charge, do you get your account back? I must admit, I just spent 10 minutes Googling Regent's Canal to see if anything interesting popped up, but there haven't been any explosions since the Tilbury barge in 1874 -- damned wild geese!
William Tanksley
"Cracks at ITER" reminds me of the math joke that after an earthquake in Menlo Park, the SLAC was renamed to the Stanford Piecewise Linear Accelerator.
What's up with finding Regent's Canal?
