Signs of the Times

https://kirstenhacker.wordpress.com/2020/01/07/signs-of-the-times-2/

by nixtaken • 7 years ago

43 59 7 years ago

kirstenhacker.wordpress.com

basicplus2• 7 years ago

userbinator• 7 years ago

..."based on what I think is more likely".

yorwba• 7 years ago

Don't forget the observations that determine the image content. "what I think is more likely" is about describing the relationship between observations and image, whereas the most likely image is produced from actual observations using that relationship to draw inferences.

rafaelvasco7• 7 years ago

Not entirely related with the article, will probably be downvoted knowing HN but the title makes something come to mind: Science is being so hold back by scientists that are too afraid to go beyond their model. Someone in a book wrote that Science is a religion too in a sense because it is as dogmatic as any religion. I tend to agree.

snovv_crash• 7 years ago

Science progresses one funeral at a time?

dsfyu404ed• 7 years ago

<subject> progresses one funeral at a time?

One of my favorite quips along those same lines is "X is Y years of progress being held back by Z years of tradition."

alexithym4• 7 years ago

It's incorrect to compare science to religion.

In science, there are no facts that are presented as being undeniably true. Rather, all scientific facts can be independently derived and understood to be true by any person. This emphasis on reproducibility means that no scientific lesson has to be taken on faith, eliminating any sense of dogma.

Furthermore, as our understanding of the world improves, we often challenge dated scientific ideas and prove them wrong using empirical evidence. If science was truly dogmatic this would never happen, because dogma requires facts to be presented as being undeniably true, forever.

SimplyUnknown• 7 years ago

If reproducibility is so important then why are we in the middle of a reproducibility crisis? [1]

If reproducibility was really that important then the scientific body would be more focused on replicating results rather racing to being the first group to publish an idea or securing funding.

Scientific foundations are seldom challenged or refuted. The last paradigm shift in physics or maths was over one hundred years ago. And as soon as someone challenges them (as the author does) the criticisms are refuted.

Say what you will, but science is mostly a social affair. Works of prominent groups are more likely to be accepted than works of groups with less status in the scientific community. Belief in scientific work may not rely on faith, but it is a far cry from pure scientific method.

I'm not trying to be cynical here and do think a good scientific body advances humanity, but the current state of the scientific community might be closer to religion than one hopes it should be.

[1] https://en.wikipedia.org/wiki/Replication_crisis

AnIdiotOnTheNet2• 7 years ago

In science, there are no facts that are presented as being undeniably true.

Not quite correct. Science and math are full of axioms.

rafaelvasco7• 7 years ago

It's absolutely correct as they both originate from the same place: The need to understand reality. They just took different approaches. Both have a lot to give to one another. I mean religion in a general sense. Those that say otherwise are too focused on one side and ignoring the other. Both of them have distortions, dogmas and bs of course.

eesmith8• 7 years ago

Except for all those times that scientists went beyond their model and overturned the previous models.

Overturned scientific models: Lyellian uniformitarianism, the pre-Agassiz model that the Earth was gradually cooling so Ice Ages couldn't have existed (see https://en.wikipedia.org/wiki/Ice_age#History_of_research ), Newton's theory of gravitation, and classical statistical mechanics (the "ultraviolet catastrophe").

Plus, the wide-spread belief that "all models are wrong" - https://en.wikipedia.org/wiki/All_models_are_wrong - or the modified "all models are wrong, but some are useful".

Which doesn't match the dogmatism attributed to religion, does it?

dnautics6• 7 years ago

Well yeah, but those are all rather old results.

empath75• 7 years ago

Newton’s theory of gravitation wasn’t tossed out; general relativity produces Newtonian gravity within certain parameters. Lots of theories are perfectly valid in a limited domain, while not being “fundamental”.

m4r35n3572• 7 years ago

I bet you can't back up a single one of your claims. Are you actually qualified to make them?

CaptArmchair• 7 years ago

What relevant measure would you use to support your claim that 'science is hold back'? And what arguments support your claim that scientists are by far the root cause to this?

'scientific progress' is a paradox. There is no absolute way of measuring the velocity at which science advances. You can't translate this into a value on a speed-o-meter and then make assertions based on what the dial says.

The entire point of science is the discovery of new knowledge and insights about reality. It's the pursuit of trying to know the unknown. Inherently, it's not possible to know up front (a) how long it will take to validate your thesis and (b) whether validating your thesis will yield the result you hoped to find (i.e. a cure for cancer)

Looking back at science in the past and then comparing that to today's state of affairs is hindsight bias.

"But, it used to be that scientists got more funding to do research. Think about the Race to the Moon!"

But then you forget that science isn't practiced in a vacuum. Never has. Scientists happen to work in earnest on particular questions because the external circumstances are aligned in a way that enables - or disenfranchises - them to work on those questions.

Older research always predicates current research. For example, there would have been little point to assume that machine learning could have been a thing in the late 1800's as neither Von Neumann nor Turing had come around to set the stage. Science begets science, and you can't rush that retroactively.

Then there's politics and economics. Full time science comes at a premium. Science happens where there's money. Either through public grants, in academic institutions or large institutions with a sizable R&D section. As such, the type of science that dominates an era depends largely on the zeitgeist.

Reducing science to pure quantitative research or the empirical method is intellectual dishonest as well. Any research is based on a foundation of need, curiosity and imagination.

As such, the number of papers published or conferences attended are very much skewed reflection of the generated value.

rafaelvasco7• 7 years ago

Sorry, 'science is held back' is the correct term. Thank you.

Psyladine2• 7 years ago

Specifically in mathematics, Grigori Perelman abandoned the field due to political favoritism by Shing-Tung Yau.

In more *systemic terms, a cabal of researchers focused effort (and more importantly, limited funding) to only one aspect of alzheimers research: https://news.ycombinator.com/item?id=21911225

Psyladine2• 7 years ago

I should mention pre-existing biases; my default setting is skeptical. I don’t really believe that black holes exist

Wonder if dark matter/energy will turn out to be the Aether of the 21st century.

pete762• 7 years ago

What does dark matter have to do with black holes?

davedx• 7 years ago

They are related topically in that yesterday new research cast the very existence of dark matter into doubt, and this blog post is kind of trying to do the same (albeit with less rigour than the actual scientific paper posted here yesterday).

mumblemumble2• 7 years ago

Well, MACHOs are one hypothesis for explaining dark matter, and they're typically considered to include (but not be limited to) black holes.

mumblemumble2• 7 years ago

I think it depends on what you mean by "dark matter" and "dark energy". If you think they're theorized substances or energy sources (which seems to be the preferred interpretation in the press), then, yeah, there seems to be a decent chance that that will be the case.

If you think, as more cautiously-minded folks seem to, that they're just clever catchwords to describe troubling lacunae in our understanding of how the universe works, not so much. There really are observations that we have trouble explaining.

burke• 7 years ago

If black holes don’t exist, what happens when gravity overpowers electromagnetism and the strong force in a region of high mass concentration? I don’t get what the alternative hypothesis could be. Do we just get arbitrarily-large neutron stars or something?

vincent-toups• 7 years ago

Something like a fuzzball:

https://en.wikipedia.org/wiki/Fuzzball_(string_theory)

But there I think most physicists think something like the solution to the Einstein field equations does obtain within most of the interior region of the objects we call black holes. They just believe that the region of the singularity may have some other physics so that it doesn't become singular.

abnry2• 7 years ago

TL;DR "I don't believe black holes exist and it is easy to p-hack your image processing algorithm, so I don't buy the media hype of the first image of a black hole."

dnautics6• 7 years ago

I didn't believe the amyloid hypothesis of Alzheimer's and I complained it was easy to p-hack all sorts of observations, and then I quit science.

ruarai• 7 years ago

This article read a little annoying. I don't know why this PhD holder failed to find the actual paper describing the algorithm: https://www.cv-foundation.org/openaccess/content_cvpr_2016/p...

The author is making it sound like the work is just p-hacking when it's /very clearly/ not.

This single paragraph in the Method section explains it best:

"Reconstructing an image using bispectrum measurements is an ill-posed problem, and as such there are an infinite number of possible images that explain the data [28]. The challenge is to find an explanation that respects our prior assumptions about the “visual” universe while still satisfying the observed data."

This is literally all of physics! You can say that Newtonian physics is 'just curve fitting' just the same.

I don't know enough about the philosophy of science to make a further argument here but clearly the author does not either.

whatshisface• 7 years ago

The allegations are that the black hole image could have turned out to be anything, given different starting assumptions about what it would look like. The exact nature of this variation, along with the degree to which the other possibilities would be plausible, is up for debate. Unfortunately that debate is very technical and not suited for general-audience blog posts.

CrazyStat2• 7 years ago

The author of the post ignores all the work that went into validating that the methods they used give reasonably consistent results, including simulation studies and multiple blind analyses to make sure that they weren't just making the image they wanted to see.

This is discussed in the Caltech talk which is embedded in the post. It seems the author cherrypicked all the quotes that sound bad and ignored the parts that disagree with her narrative.

raverbashing• 7 years ago

Yeah this article is clickbait

I would paraphrase that "saying" about teaching as thus: "those who can't do, nitpick"

The TED talk was for a different algorithmic technique.

Couple that with assertions like "An image with a hole or without a hole looks the same in the amplitude domain." and I see that this article has as much credibility as climate deniers.

Same thing with the Ligo "naysayers" who tried to disprove the experiments without knowing the basics of signal processing using FFT, etc

abnry2• 7 years ago

The reason the article is annoying is that there could be some very interesting criticisms based on what was written. "P-hacking" and related statistical fudging is a valid concern, but you need technical details. The skepticism towards black holes because of skepticism towards singularities is an interesting idea. As such, it just seems like naysayer fluff.

dzdt• 7 years ago

Its ironic.

The author formed a theory that the black hole image research was flawed, specifically that the researchers are filtering for images with the right shape and merging these to produce an image similar to predictions. The author then looks at various talks and presentations and filters for isolated statements that out-of-context support the idea of the research flaws.

The author has a flawed approach exactly analagous to the accused flaw in the black hole research.

dnautics6• 7 years ago

In one case data directly about the object in question were thrown out, in the other case, data were ignored (analogously: throwing out images of stars when cataloging galaxies based on priors of galaxy spectral properties vs star spectral properties)

iamthepieman• 7 years ago

so if the data could represent any image (not sure if the term "any" is truly infinite or just very large) then at least Bouman's algorithm and work sets a stake in the ground and challenges other scientists to explain why it does or doesn't belong there.

Reading this article makes it seem like the work is a scientific version of my kids artwork. Interesting to the right audience, tells you a lot about who created it but bearing only a slight semblance to reality.

CrazyStat2• 7 years ago

If you watch the Caltech talk (embedded in the blog post) or read the original published paper, they did quite a bit of work to validate that their method was actually giving them good results and not just what they wanted to see. The author of the post doesn't mention any of this, presumably because it doesn't fit her narrative.

losvedir• 7 years ago

I think this article is running off on a wild goose chase.

I remember looking into this when it all came out and I don't think Prof Bouman's algorithm was actually used for that picture. Her TED talk was great, and I think that's why everyone latched onto it, but IIRC they ended up using a different reconstruction algorithm.

Some people were unreasonably concerned with who actually did what with that image, and I wouldn't normally care about it. It's just that this article is trying to cast doubt on the existence of black holes by attacking an algorithm that wasn't even used to generate the picture they're showing on the page.

raxxorrax• 7 years ago

Without phase information the cat probably looks like a black hole if transformed back into image space.

Signs of the Times

Related Stories