Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

Comment author: polymathwannabe 16 January 2015 08:26:16PM 1 point [-]

The article you cite says,

This study does not dispute the safety of vaccines but reinforces the need to study long-term effects of early exposure to neuro-toxic substances on the developing brain.

The toxicity of Al is much lower than that of thimerosal

Mild post-vaccine symptoms in young infants, especially neonates, are non-specific and considered tolerable; rare (neurologic) adverse effects are unlikely to occur as a result of adjuvant-Al per se or in combination with thimerosal-Hg.

From the article I didn't gather what type of exposure was more worrying to the authors---acute or chronic. They seem to admit acute exposure has been proven safe, but on the other hand they dare not make any definite statements on chronic exposure. And above all, they never suggest that vaccinations should be stopped: in their conclusion they make it very clear that the purpose of understanding better the toxicity profile of vaccines is for increasing trust in vaccination.

Comment author: PhilGoetz 22 March 2015 08:36:17PM 0 points [-]

in their conclusion they make it very clear that the purpose of understanding better the toxicity profile of vaccines is for increasing trust in vaccination.

That's a huge red flag right there. It means they've already decided what research must prove.

Comment author: PhilGoetz 11 March 2015 06:17:38PM *  3 points [-]

Basic question about bits of evidence vs. bits of information:

I want to know the value of a random bit. I'm collecting evidence about the value of this bit.

First off, it seems weird to say "I have 33 bits of evidence that this bit is a 1." What is a bit of evidence, if it takes an infinite number of bits of evidence to get 1 bit of information?

Second, each bit of evidence gives you a likelihood multiplier of 2. E.g., a piece of evidence that says the likelihood is 4:1 that the bit is a 1 gives you 2 bits of evidence about the value of that bit. Independent evidence that says the likelihood is 2:1 gives you 1 bit of evidence.

But that means a one-bit evidence-giver is someone who is right 2/3 of the time. Why 2/3?

Finally, if you knew nothing about the bit, and had the probability distribution Q = (P(1)=.5, P(0)=.5), and a one-bit evidence giver gave you 1 bit saying it was a 1, you now have the distribution P = (2/3, 1/3). The KL divergence of Q from P (log base 2) is only 0.0817, so it looks like you've gained .08 bits of information from your 1 bit of evidence. ???

Comment author: PhilGoetz 11 March 2015 11:22:25PM *  2 points [-]

I think I was wrong to say that 1 bit evidence = likelihood multiplier of 2.

IF you have a signal S, and P(x|S) = 1 while P(x|~S) = .5, then the likelihood multiplier is 2 and you get 1 bit of information, as computed by KL-divergence. That signal did in fact require an infinite amount of evidence to make P(x|S) = 1, I think, so it's a theoretical signal found only in math problems, like a frictionless surface in physics.

If you have a signal S, and P(x|S) = .5 while P(x|~S) = .25, then the likelihood multiplier is 2, but you get only .2075 bits of information.

There's a discussion of a similar question on stats.stackexchange.com . It appears that the sum, over a series of observations x, of

log(likelihood ratio = P(x | model 2) / P(x | model 1))

approximates the information gain from changing from model 1 to model 2, but not on a term-by-term basis. The approximation relies on the frequency of the observations in the entire observation series being drawn from a distribution close to model 2.

Comment author: PhilGoetz 11 March 2015 06:17:38PM *  3 points [-]

Basic question about bits of evidence vs. bits of information:

I want to know the value of a random bit. I'm collecting evidence about the value of this bit.

First off, it seems weird to say "I have 33 bits of evidence that this bit is a 1." What is a bit of evidence, if it takes an infinite number of bits of evidence to get 1 bit of information?

Second, each bit of evidence gives you a likelihood multiplier of 2. E.g., a piece of evidence that says the likelihood is 4:1 that the bit is a 1 gives you 2 bits of evidence about the value of that bit. Independent evidence that says the likelihood is 2:1 gives you 1 bit of evidence.

But that means a one-bit evidence-giver is someone who is right 2/3 of the time. Why 2/3?

Finally, if you knew nothing about the bit, and had the probability distribution Q = (P(1)=.5, P(0)=.5), and a one-bit evidence giver gave you 1 bit saying it was a 1, you now have the distribution P = (2/3, 1/3). The KL divergence of Q from P (log base 2) is only 0.0817, so it looks like you've gained .08 bits of information from your 1 bit of evidence. ???

Comment author: mwengler 24 February 2015 02:31:18PM 1 point [-]

That's the goal. What, you want there to be humans a million years from now?

Is that true, or are you just being cleverly sarcastic? If that is the goal of CEV, could you point me to something written up on CEV where I might see this aspect of it?

Comment author: PhilGoetz 11 March 2015 04:41:15PM *  0 points [-]

I mean, that's the goal of anyone with morals like mine, rather than just nepotism.

Comment author: Luke_A_Somers 23 February 2015 11:47:52AM *  1 point [-]

I'm talking about a real and important distinction, which is the degree of freedom in values to give the next generation. Under standard CEV, it's zero.

No, it's not.

Zero is the number of degrees of freedom in the AI's utility function. not the next generation's utility functions.

Comment author: PhilGoetz 11 March 2015 04:37:25PM 0 points [-]

When using the parent-child relationship as an instance of CEV, it is. The child takes the position of the AI.

Comment author: Luke_A_Somers 21 February 2015 03:36:46AM *  1 point [-]

If we value them getting to go and make their own choices, then that will be included in CEV.

If we do not value them being brainwashed, it will not be included in CEV.

I strongly suspect that both of these are the case.

Comment author: PhilGoetz 23 February 2015 06:33:19AM *  2 points [-]

I know that is the standard answer. I tried to discourage people from making it by saying, in the parent comment,

I know somebody's going to say, "Well, then that's your utility function!"

I'm talking about a real and important distinction, which is the degree of freedom in values to give the next generation. Under standard CEV, it's zero.

I don't think that parameter, the degree of freedom, should be thought of as a value, which we can plug any number we like into. It should be thought of as a parameter of the system, which has a predictable impact on the efficacy of the CEV system regardless of what values it is implementing.

I don't think people allow their children freedom to make up their own minds because they value them doing so. They do it because we have centuries of experience showing that zero-freedom CEV doesn't work. The oft-attempted process of getting kids to hold the same values as their parents, just modified for the new environment, always turns out badly.

Comment author: Toggle 21 February 2015 05:06:22AM *  3 points [-]

You're comparing two different measures, military casualties versus total casualties. For a reasonably good apples-to-apples measure of wartime violence over time, Better Angels of our Nature is one that I found readable and informative.

IIRC, overall per capita violent death during the world wars was roughly comparable to living in a tribal society during a typical time- of course, for us, those were the exception and not the rule.

Comment author: PhilGoetz 23 February 2015 06:09:15AM -1 points [-]

I am deliberately looking at military casualties, to highlight that tribal "military" casualties couldn't possibly be that high. I would guess that the fraction of deaths that were civilian was higher in both world wars than in tribal conflicts. Tribal conflicts are, AFAIK, almost always strictly men killing other men. Per capita comparison is distorted by the longer lifespan of people in the 20th century. Just having more people live past the age of 40 shouldn't, I'd think, make your age look more peaceful.

Comment author: Tom_McCabe 01 September 2007 11:18:50PM 18 points [-]

"You can actually give a semi-plausible justification of special relativity based on what was known in 1901."

You can give a semi-plausible justification for anything. It was obvious at the time that our knowledge was incomplete, but the specific *way* in which our knowledge was incomplete was still a mystery. It is very easy to invent a plausible-sounding quack theory of physics; that is why we have the Crackpot Index.

Comment author: PhilGoetz 23 February 2015 05:49:59AM *  1 point [-]

Well, the key equation of special relativity had already been written down in 1901 as the Lorentz contraction. People just hadn't thought of interpreting it so literally.

But the larger point is still valid. Change the date to 1894 and it would be a complete novelty.

Comment author: fowlertm 21 February 2015 03:48:19PM 1 point [-]

Why? What's wrong with wanting to be masculine?

Comment author: PhilGoetz 23 February 2015 05:37:44AM *  -1 points [-]

If it were wrong, it would be a problem, not problematic. That defies the dictionary definition, but "problem" can mean something with a simple solution that hasn't yet been implemented, while "problematic" connotes a persistent problem with no easy solution.

The difficulties with it are already listed in the post, as they're the motivation for the post. Though it might be more fair to say gender is problematic.

Comment author: PhilGoetz 21 February 2015 01:20:13AM 52 points [-]

Eliezer and I are now part of the literary canon.

At least, we're both taught in the English department at Princeton. Anne Jamison's course, "Fanfiction: Transformative works from Shakespeare to Sherlock", will cover Eliezer's Harry Potter and the Methods of Rationality on March 2, and one of my short stories, "The Magician and the Detective", on March 4.

View more: Next