10 ms·
I am very skeptical of the Event Horizon Telescope (EHT) images because they are not following a scientific method that results in a true representative image o
by fasteddie31003 4y ago
I am very skeptical of the Event Horizon Telescope (EHT) images because they are not following a scientific method that results in a true representative image of their target. In my opinion astronomy is jumping the shark with these images by making this a big PR stunt.
I've looked at their methods for their earlier images and they seem to be hunting for a circle that looks like a black hole in their data. The EHT's full imaging stack has never been calibrated by looking at a known celestial body to compare images to validate their algorithms. They have calibrated their signals from results of other instruments, but their imaging algorithms change to fit their wanted results. This is my biggest problem with their approach. Anyone can modify algorithms of any arbitrary data to get an image of a glowing circle. A better method that shows a more true image would be to calibrate their imaging algorithms against a known celestial body to make sure their techniques produced comparable results from other instruments. Then they should have taken their calibrated imaging algorithms and gave it data from their target.
I'd have more confidence in the EHT if they would not change their imaging algorithms across images and give a side-by-side comparison of a known celestial body that other radio telescopes have imaged to verify their whole imaging stack.
To me this a just a big PR stunt and I'm very skeptical of their image.
- wumpus 4y agoDidn’t I already reply to you weeks ago about a similar comment? Now you can read the actual papers which explain in detail why your comment is incorrect.
- fasteddie31003 4y agoI did read them and I'm not satisfied with their methods.
- wumpus 4y agoYou read the 10 just published papers? Impressive.
- Mehringotio 4y agoEither you are an expert in this field than it should be you who writes a clear argument and present it or you are not an expert. Aren't they are peer reviewed? What is your expertise in astrophysics?
- gilbetron 4y agoHe has a theoretical degree in astrophysics.
- Saint_Genet 4y agoHe still didn’t read and understand 10 papers this fast
- gilbetron 4y agoit was a joke from a Fallout game - he has a "theoretical degree in astrophysics" not a "degree in theoretical astrophysics" :)
- ajcp 4y agoWelcome aboard!
- wumpus 4y agoHow about this independent analysis: https://ui.adsabs.harvard.edu/abs/2022arXiv220510267P/abstract https://ui.adsabs.harvard.edu/abs/2022arXiv220510267P/abstra...
- belter 4y agoI think you are talking about the conversation we had here. So in this case you replied to a different person. https://news.ycombinator.com/item?id=31281773 https://news.ycombinator.com/item?id=31281773
- panda-giddiness 4y agoThanks for calling them out (again). Honestly people in this forum are too eager to upvote contrarian opinions. I hold a PhD in physics, and I wouldn't feel qualified to challenge ETH's work.
- wumpus 4y agoI’m fine with questions and challenges, but not aggressive ones from people who don’t like my explanation and will not read the literature.
- aortega 4y agoScience is not religion and contrarian opinions strengthen science results. I think it's healthy to always have a dose of skepticism, and is not uncommon for astrophysics papers to be retracted or refuted.
- panda-giddiness 4y agoAbsolutely, but those retractions are led by other astronomers/astrophysicists, not lay people presenting straw men. I think people's priors here are out of whack -- perhaps the truth does lie somewhere in the middle, but if so, it will be much closer to ETH's version than the OP's.
- belter 4y agoAstronomers/astrophysicists started first, by misrepresenting the process and bringing the talk of taking photos and having "visual evidence". "...We have seen, and taken a picture, of a Black Hole." https://youtu.be/lnJi0Jy692w?t=456 https://youtu.be/lnJi0Jy692w?t=456
- wumpus 4y agoYes? He’s correct.
- 4y ago
- malfist 4y agoWhy do you think you know more about astronomy than a team of scientists publishing peer reviewed information?
- fasteddie31003 4y agoPeer reviewed has nothing to do with the scientific method. It is just a layer of scientific bureaucracy. Just give me a calibration image of another celestial body and I'll be happy.
- zenmaster10665 4y agothis guy astronomies.
- malfist 4y agoYou realize peer review isn't just something you do and they automatically publish you, right? Peer review makes sure your science is sound. But I'm sure you're more of an expert than anybody serving in a peer review board.
- fasteddie31003 4y agoPeer review has let plenty of bad science through. Don't burn me at the stake because I speak heresy against the holy theologians of astronomy.
- DiogenesKynikos 4y agoBad papers do get through, but anything as high-profile as the first image of the supermassive black hole at the center of the Milky Way will get intense scrutiny.
- malfist 4y agoAll you're saying is: They're wrong, not gonna say how, but trust me. Why should I trust you that you know more about astronomy than a team of astronomers and a peer review board? I've asked you this three times so far and you've not given an answer.
- h2odragon 4y agoI loved to pick on my astrophysicain buddies about their over-reliance on interpolation. It's not invalid in itself but they do seem fond of making sweeping generalizations that fit neatly into the smooth spots in their spectrums.
- dotnet00 4y agoChecking against callibration measurements isn't the only way to do things. With M87's SMBH for instance, they had a simulated image of what they ought to see given general relativity and what they knew about M87. The extracted data matches that fairly closely, thus lending credibility to the result for obtaining parity with physics without actually applying those physics to the data.
- JacobThreeThree 4y ago>With M87's SMBH for instance, they had a simulated image of what they ought to see given general relativity and what they knew about M87. How confident are we that our current simulations accurately reproduce the universe? Given how many unpredicted and supposedly impossible exoplanet and star configurations that keep being found, I'd say the current model is not doing so well on the prediction front. https://www.science.org/content/article/forbidden-planets-understanding-alien-worlds-once-thought-impossible https://www.science.org/content/article/forbidden-planets-un...
- dotnet00 4y agoThose configurations aren't "impossible" due to fundamental well tested physics like general relativity. They're "impossible" from the perspective of our understanding of the formation of planetary systems, which is understandably less well developed given how much more difficult it is to study since as a science, exoplanet detection is just 30 years old, with the majority of detections being less than 15 years old. In comparison, general relativity is one of the most well tested theories in physics, having been undergoing rigorous testing for over a century, and regardless, the point stands that the results were close to the model, thus providing evidence for the model's validity.
- JacobThreeThree 4y agoIf general relativity based models have struggled to help predict or explain planetary system formation and exoplanet observations, I think we should have the same level of expectation about how helpful a general relativity based black hole model will be.
- photochemsyn 4y agoYou mean, this? https://iopscience.iop.org/article/10.3847/2041-8213/ac6615 https://iopscience.iop.org/article/10.3847/2041-8213/ac6615 >"Since the interferometric measurements are often incomplete in the Fourier domain, the inverse problem of reconstructing an image from the observed data set is usually underdetermined. Consequently, the image reconstruction requires prior information, assumptions, or constraints to derive a reasonable image from the infinite number of possibilities that can explain the measurements." They seem to have gone to great lengths to address this issue, however. Multiple imaging approaches, synthetic data tests, etc. https://iopscience.iop.org/article/10.3847/2041-8213/ab0e85/meta https://iopscience.iop.org/article/10.3847/2041-8213/ab0e85/... Why would looking at something that isn't a supermassive black hole at the center of a galaxy prove that this approach works or doesn't work? You'd have different constraints to apply. See for example VLBI measurements of quasars, which seem to employ the same kind of imaging approach. Theoretical models of quasars aren't the same as theoretical models of black holes. They seem to be using these theoretical models as constraints on data interpretation. Unless the theory is completely wrong, which seems unlikely, this looks like a valid approach. https://arxiv.org/pdf/1701.04760.pdf https://arxiv.org/pdf/1701.04760.pdf
- queuebert 4y ago>"Since the interferometric measurements are often incomplete in the Fourier domain, the inverse problem of reconstructing an image from the observed data set is usually underdetermined. Consequently, the image reconstruction requires prior information, assumptions, or constraints to derive a reasonable image from the infinite number of possibilities that can explain the measurements." If the original commenter has an issue with this, wait 'til they find out about modern CT scans. Or hell even JPEGs.
- heyitsguay 4y agoIt's the usual HN physics thread stuff - for some reason physics really brings out the confidently incorrect crowd.
- oldgradstudent 4y agoFunny that you mention CT scans. To quote from a paper[1] on the history of CT scanning: "Initial images were of inert objects, then specimens from an abattoir, including bullocks brains and pigs bodies" or "The prototype was installed at Atkinson Morley’s Hospital in South London where the first patient, a middle aged lady with a suspected frontal lobe tumour, was scanned on 1st October 1971. The surgeon who operated on her shortly afterwards reported that ‘‘it looks exactly like the picture’’" The default scientific approach to a new imaging algorithm, and especially a new implementation, it to try it on a simple well understood examples first. [1] Beckmann, E. C. (2006). CT scanning the early days. https://www.birpublications.org/doi/10.1259/bjr/29444122 https://www.birpublications.org/doi/10.1259/bjr/29444122
- jankovicsandras 4y agoI share your skepticism. To others: Before voting me down, please consider these arguments. As I understand, this image is not from visible light, not a photo, more like plotting radio measurements. "black holes" are the brightest objects in the universe. It's a bit like an alien scientist showing an X-ray picture of a skeleton:"This is a human!" Some quotes from https://en.m.wikipedia.org/wiki/Black_hole https://en.m.wikipedia.org/wiki/Black_hole : "Moreover, quantum field theory in curved spacetime predicts that event horizons emit Hawking radiation, with the same spectrum as a black body of a temperature inversely proportional to its mass. This temperature is of the order of billionths of a kelvin for stellar black holes, making it essentially impossible to observe directly." The same article is full of words like "implies", " inferred", "indirect" etc. It's not directly proven black holes even exist, see Alternatives paragraph. I'm not sure it's physically possible to take a visible light picture of a "naked" event horizon - then black hole images are pure phantasy.
- sailingparrot 4y agoWhile the skepticism of the comment you are replying to is understandable, yours is not. "Visible light" is not a better or truer electromagnetic spectrum than x-ray or radio. It's a human construct only defined by what our human eyes have evolved to capture, based on what is most useful for us here on earth, nothing more. There is nothing wrong with getting measurement of an object in any random frequency that happens to be the most interesting/practical. Rendering those results as an image in visible light is also simply the best way to visualize the results for us humans. It is certainly more understandable than hundreds of plots. > It's a bit like an alien scientist showing an X-ray picture of a skeleton If that alien organism has evolved to see x-rays, as it might be the most useful spectrum for their particular environment, they would find your comment quite puzzling as that would be exactly what they would see should they meet us.
- gnulinux 4y ago> It's a bit like an alien scientist showing an X-ray picture of a skeleton:"This is a human!" What is the problem here? Surely skeleton (or other things detectable via X-rays) is an integral part of a human. If you have no way of seeing an alien, but have an x-ray of them, wouldn't you say we have a significantly better understanding of this alien species, than having nothing at all?
- gamblor956 4y agoOne can always rely on some commenter on HN assuming that professionals in another industry don't know how to do their job properly and just need some random programmer to tell them how they should actually be doing things. I happen to live and work near JPL. The astronomers there have none of the qualms you do about this image, because it turns out the professionals working on this image actually took such things you're complaining about into consideration when refining the data. And if you had read the publication accompanying this image, you would have known that. Also, they weren't just looking for a glowing circle...That's a shockingly ignorant way to characterize their work.
- queuebert 4y agoThe point spread function of the detector is pretty well characterized, and from there the image is basically just a Fourier transform. It's not as magical as you might think. Then it looks like they average a bunch of these images to get a maximum likelihood image. What is your issue with that? Besides, to calibrate on a known celestial body, they'd have to have another micro-arcsecond resolution radio telescope to use. Do you have one they can borrow?
- fasteddie31003 4y agoCould they not zoom out and take a picture of the stars around Sagittarius A* to compare that to the knowns locations of these stars?
- jasonwatkinspdx 4y agoThey absolutely could, but doing so would be a pointless waste of nontrivial resources.
- m_mueller 4y agoThat doesn’t seem a waste to me. Use the same algorithm, data pipeline and sensors to take imagines of some far away galaxies that are well imaged already, see what you get. If it matches it will certainly be a good argument against the doubters.
- quasarj 4y agoThank you! I read over their process from the first one and it sounded like they were just using ML to make up most of the image! I need to read a lot more of the details but I am quite skeptical of these images...
- fsh 4y agoThe papers do not mention any machine learning.