Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Bayes's Theorem: What's the Big Deal?

blogs.scientificamerican.com

Post reply on HN
Print view View original

181–190 of 267 posts

Re: Bayes's Theorem: What's the Big Deal?

#181

Post by URSpider94 »

Jabbles wrote: post #67
Anderkent wrote: post #50

Earlier quoted context omitted.

Doctors observe a result of the test, and know the basic probabilities (in the example, 99% test accuracy, 1% of population have the disease). The problem is that they [often] draw incorrect conclusions from those observations (99% test accuracy and you tested positive? well then you likely - 99% - have the disease, right?). The question formed as 'your one patient tested positively' is more immediately relevant, I'd…

Of course, doctors do not randomly assign tests to patients. Their prior that a patient has a disease is a lot higher than the background frequency of it occurring. Getting them to estimate their prior would be interesting.

Except when they do. For example, when 100% of men above a certain age are screened for prostate cancer, or 100% of women above a certain age are screened for breast cancer. Both cases spawned major public health campaigns to encourage screening, followed years later by recommendations AGAINST 100% screening, based on the high degree of false positives and unnecessary treatment.

Other cases that come to mind: -- doctors who offer "full body scans" as a part of an executive physical; you're pretty much guaranteed to turn up something that is 2 sigma away from the population norm, somewhere in the body, on such a scan

-- spinal x-rays for back pain. Doctors almost always find something abnormal, and use that to justify the back pain and treat aggressively. But, we don't really have a good prior; if you x-rayed 1000 people off the street, would we find similar abnormalities frequently?

Re: Bayes's Theorem: What's the Big Deal?

#182

Post by conjectures »

IanCal wrote: post #31
Laaw wrote: post #14

I've been saying this for years , and this is a large reason why I find the LessWrong folks to be almost entirely full of it. Their inability to come up with accurate priors is completely lost on many of the folks who follow this kind of thinking. A couple of comments are saying, "no duh" to this article, but those folks likely don't realize quite how many other people are falling into this trap. "Garbage in, garbage…

Do priors just start you off closer to the truth? That is to say, if you start with any prior, will enough additional pieces of evidence always let you converge on the truth? Does anyone commonly set their priors to be a distribution? Perhaps a range or actually a normal distribution to represent a prior with uncertainty?

Yes, but.

The technical term for the condition you're interested in is consistency. There is a theorem (Doob) regarding the consistency of Bayes estimates. [1]

In simple cases it means that under any non-silly prior your posterior will converge to the truth. E.g. mean of a normal distribution with a normal prior.

The complex cases include things like models that expand when given more data.

[1] http://www4.stat.ncsu.edu/~sghosal/papers/bhurev.pdf

Re: Bayes's Theorem: What's the Big Deal?

#183

Post by Nostripus »

RodericDay wrote: post #78
TeMPOraL wrote: post #62

Earlier quoted context omitted.

I'm not that huge advocate of LW. I used to frequent the site, I read the Sequences and generally liked what I saw. RE parts you quoted. Of course they look bad, because they've been ripped out from context. Race: http://lesswrong.com/lw/kk/why_are_individual_iq_differences... This is an article asking, to quote: "But why is it that the rest of the world seems to think that individual genetic differences are okay, wh…

I guess I just disagree with you that they're taken out of context? 1. Bond posits Yudkowski thinks that racism is mostly due to genetic difference, and not about the deliberate, mostly political disenfranchisement of minorities. You say this is a wrong because his appraisal of the situation as being caused by genetic disparities is benevolent, and that he wants to work to "fix" it. These aren't in contradiction- to…

>The whole point is that these kind of people find these questions ("is rape really bad?") really intriguing rather than obvious.

That isn't the question the story is asking. Second of all, Eliezer says:

>This is a work of fiction. In real life, continuing to attempt to have sex with someone after they say 'no' and before they say 'yes', whether or not they offer forceful resistance and whether or not any visible injury occurs, is (in the USA) defined as rape and considered a federal felony. I agree with and support that this is the correct place for society to draw the line. Some people have worked out a safeword system in which they explicitly and verbally agree, with each other or on a signed form, that 'no' doesn't mean stop but e.g. 'red' or 'safeword' does mean stop. I agree with and support this as carving out a safe exception whose existence does not endanger innocent bystanders. If either of these statements come to you as a surprise then you should look stuff up. Thank you and remember, your safeword should be at least 10 characters and contain a mixture of letters and numbers. We now return you to your regularly scheduled reading. Yours, the author.

The point that the protagonists legalized rape, is to demonstrate that the protagonists morals are just as far from us, as the baby eaters or super happies. That the story isn't about two alien societies, but three.

Second of all, what "these kind of people" find intriguing is that our morals are for the most part due to our circumstances. That in the future, people may very well accept what we find today to be abhorrent, and find abhorrent what we accept.

That is the point of "is rape bad?". Because it is simply an example of something that is obviously bad. Just like some time ago there were other things that were considered obviously bad, that today we accept (same sex relationships for example). And if we are aware that the future will look back and mock us for our obviously wrong morals, then maybe we will be faster to get on the ball with the next gay rights or what ever.

Are there moral beliefs you hold today, that you think the future might condemn you for? Or do you think in this time and place you've acquired a perfect set of morals?

Re: Bayes's Theorem: What's the Big Deal?

#184

Post by conjectures »

cronjobber wrote: post #34

Here's a proposal: Bayesian scientists shouldn't select their own prior. Instead publish how your results would update any prior, including the one picked by me, the reader. I certainly haven't thought this through, but maybe this would make science more modular: combine the updates from M studies and calculate the new, combined update. Statisticians, does this work?

On the second point. It's called meta-analysis and is an important area of research. However it's very far from operating automagically and requires substantial manual input.

On the first point this is equivalent to contracting a builder to build a house in any style.

Choosing a prior isn't always just about picking a wide band around 3 or a narrow band around 2. This is like the builder offering a choice of curtain colours from a swatch.

Making a commitment to implement any prior could be totally. Like the overcommitted builder being asked to reconstruct R'lyeh.

Re: Bayes's Theorem: What's the Big Deal?

#185

Post by RodericDay »

edanm wrote: post #153
RodericDay wrote: post #78

Earlier quoted context omitted.

I guess I just disagree with you that they're taken out of context? 1. Bond posits Yudkowski thinks that racism is mostly due to genetic difference, and not about the deliberate, mostly political disenfranchisement of minorities. You say this is a wrong because his appraisal of the situation as being caused by genetic disparities is benevolent, and that he wants to work to "fix" it. These aren't in contradiction- to…

"1. Bond posits Yudkowski thinks that racism is mostly due to genetic difference, and not about the deliberate, mostly political disenfranchisement of minorities." You don't need to take the author's word for it. Look at the article itself. The article is not about the question of whether race affects intelligence. It's saying that that question is much less important than the fact that individuals have different IQ…

I'm not gonna do a good job at explaining this, because it's far from my subject area, but the problem is that your demand for "the relevant part" asks for blunt evidence of a subtle phenomenon.

When we talk about politics, context matters. Someone like Yudkowsky doesn't believe stupidly racist things, like a KKK member or turn-of-century politician, or even modern anti-immigration people. Bond's targeting something much more subtle.

The point is that here you have a man who naively keeps trying to push the dialogue "beyond race" ("forget race"), in a forum where if you scroll down to the comment section, you'll see Jeff H. with 5 upvotes defending Watson's racist remarks, with Epiphany at 0 upvotes talking about the cultural reasons why IQ tests fail.

It's about the framework, and what it allows, and about what allows the people championing it to be naive or indifferent about what it allows. Someone discussing coldly and clinically the pros and cons of rape isn't a rapist, but there's something else about someone who can have discussion coldly and clinically. Some people take pride in their level-headedness about tough topics, others would take it as a signal of their lack of empathy with the victims, and the signal this sends to people who do feel passionately about it.

And so Yudkowsky will discuss race in this goofy aloof kind of way- "maybe black people are intellectually inferior, let's talk about it, does it even matter in the end?", and some people will believe that this is the right kind of coolness to produce neat solutions, and others will believe it robs the steam off of the emotional connection that would engage people in political action for change.

You have to suss this kind of thing out, and if you plain don't want to believe me that there's something enabling towards racism and sexism and other various supremacist ideologies in the LessWrong ecosystem, you don't really have to.

Re: Bayes's Theorem: What's the Big Deal?

#186

Post by senekerim »

RodericDay wrote: post #30

I think Stephen Bond did some excellent takedowns of the identity politics that has arisen around Bayes' Theorem back in the day. I wonder where he's at these days. The Cult of Bayes' Theorem http://laurencetennant.com/bonds/cultofbayes.html > One of Yudkowsky's constant refrains, appropriating language from Frank Herbert's Dune, is "Politics is the Mind-killer". Under this rallying cry, Lesswrong insiders attempt to…

This is written by a person who doesn't know what he's talking about. Any time people diss Bayesian stats and name Yudkowsky in their argument (a controversial guy without a high school degree), they haven't got a clue what they are saying.

Bayesian stats is a solid mathematical theory. It is used everywhere in academia and industry, is the foundation of a lot of AI & Machine Learning research, powers many technologies that, among other things, allow you to locate this thread on the Internet and post that ignorant link here.

Please stop promoting BS.

Re: Bayes's Theorem: What's the Big Deal?

#187

Post by jameshart »

edanm wrote: post #157
jacobolus wrote: post #109

Earlier quoted context omitted.

The LessWrong folks aren’t obviously better or worse at calculating priors than anyone else. The “problem” is that their hobby is spending their free time considering outlandish scenarios, inventing arbitrary assumptions related to such scenarios, drawing questionable conclusions, and then convincing themselves that because they used logic and math, their analysis must be correct. Plenty of other folks who spend time…

So I'm a LessWronger and know a bit about the "movement", and think you are misunderstanding what "LessWrongers think". Obviously not all LessWrongers think the same thing at all, but I'm talking about the average position of the people who believe AI safety should be worked towards. I'd love to explain the basic position, and tell me where you disagree with it. This is the basic position: 1. Intelligence can be crea…

Yeah, you go off the rails around step 5. "Something being much more intelligent than us means that, in effect, it has almost absolute power over what happens in the world" makes no sense. Since when does intelligence get you power? Are the smartest people you know also in positions of power? Are the most powerful people all highly intelligent?

"whatever the goals of the AI will be, it will achieve them". Dude, if intelligence meant you could achieve your goals, Hacker News would be a much less whiny place.

Re: Bayes's Theorem: What's the Big Deal?

#188

Post by philh »

Apocryphon wrote: post #177
unimpressive wrote: post #146

Earlier quoted context omitted.

https://en.wikipedia.org/wiki/Pascals_wager I don't think it would be fair to malign philosophers because they come up with outlandish scary scenarios that scare people with OCD sometimes. It's not like LW gives Roko significant air time or serious treatment (EY freaking out and deleting it was partially principle of the thing, partially the fear that somebody might follow this road of thought to come up with somethi…

It's less about the plausibility of the thought experiment, and more about typical online drama and hysteria that ensued, which sort of belies that LW is made up of mortals like you and me. They aren't hyper-rational machines, after all.

In that case, how is it a reply to woodchuck?

"The subject matter that LessWrong considers is certainly unusual, but that alone should not be enough to call it arbitrary, questionable or outlandish."

"But let me tell you about the time LW experienced internet drama."

Doesn't seem particularly germane?

Re: Bayes's Theorem: What's the Big Deal?

#189

Post by jacquesm »

TeMPOraL wrote: post #111
jacobolus wrote: post #109

Earlier quoted context omitted.

The LessWrong folks aren’t obviously better or worse at calculating priors than anyone else. The “problem” is that their hobby is spending their free time considering outlandish scenarios, inventing arbitrary assumptions related to such scenarios, drawing questionable conclusions, and then convincing themselves that because they used logic and math, their analysis must be correct. Plenty of other folks who spend time…

Oh come on. That's how you're supposed to use math. To aid your thinking. Without it, and considering "outlandish scenarios", we would not have any scientific progress. Also, this is one strange thing - any time someone asks us (the STEM crowd), "what will I ever use math for in my life?", the default answer seems to be, "it's about having more tools for thinking, and greater clarity of thought; it'll make you smarte…

The more you extrapolate the more frequently you will need to adjust your future predictions because of errors in your initial measurements. This goes for any kind of extrapolation (for instance: plotting a course on a map), but it goes even more for extrapolating the future from limited evidence present today. Your 'best guess' might be off by many orders of magnitude if the evidence you have today is only loosely related to the future in terms of importance and where evidence may not be nearly as independent in nature as you currently perceive.

This can lead to your best guess based on available evidence being about as good at predicting the future as randomness in spite of all the apparent effort at making the predictions mathematically sound.

Re: Bayes's Theorem: What's the Big Deal?

#190

Post by philh »

jameshart wrote: post #187
edanm wrote: post #157

Earlier quoted context omitted.

So I'm a LessWronger and know a bit about the "movement", and think you are misunderstanding what "LessWrongers think". Obviously not all LessWrongers think the same thing at all, but I'm talking about the average position of the people who believe AI safety should be worked towards. I'd love to explain the basic position, and tell me where you disagree with it. This is the basic position: 1. Intelligence can be crea…

Yeah, you go off the rails around step 5. "Something being much more intelligent than us means that, in effect, it has almost absolute power over what happens in the world" makes no sense. Since when does intelligence get you power? Are the smartest people you know also in positions of power? Are the most powerful people all highly intelligent? "whatever the goals of the AI will be, it will achieve them". Dude, if in…

You're not engaging with the claim as it was meant. In context, no human being has ever been "much more intelligent" than me. Not in the same way that I am "much more intelligent" than the monkey von Neumann.

You might decide that this means edanm goes off the rails at step four, instead. But you should at least understand where you disagree.

Post reply on HN