ksvanhorn

Karma: 509

ksvanhorn 18 Jan 2011 19:35 UTC
40 points
0
in reply to: poke’s comment on: Is That Your True Rejection?
“There’s two things you get from a formal education: one is broad, you’re exposed to a variety of subject matter that you’re unlikely to encounter as an autodidact;”

As someone who has a Ph.D., I have to disagree here. Most of my own breadth of knowledge has come from pursuing topics on my own initiative outside of the classroom, simply because they interested me or because they seemed likely to help me solve some problem I was working on. In fact, as a grad student, most of the things I needed to learn weren’t being taught in any of the classes available to me.

The choice isn’t between being an autodidact or getting a Ph.D.; I don’t think you can really earn the latter unless you have the skills of the former.

Bayesian probability theory as extended logic—a new result

ksvanhorn6 Jul 2017 19:14 UTC

37 points

40 comments4 min readLW link

For-Profit Rationality Training

ksvanhorn28 Dec 2011 21:42 UTC

31 points

39 comments3 min readLW link

ksvanhorn 22 Jan 2011 6:06 UTC
24 points
0
in reply to: GBM’s comment on: Probability is in the Mind
“Is it not true that when one rolls a fair 1d6, there is an actual ¹⁄₆ probability of getting any one specific value?”

No. The unpredictability of a die roll or coin flip is not due to any inherent physical property of the objects; it is simply due to lack of information. Even with quantum uncertainty, you could predict the result of a coin flip or die roll with high accuracy if you had precise enough measurements of the initial conditions.

Let’s look at the simpler case of the coin flip. As Jaynes explains it, consider the phase space for the coin’s motion at the moment it leaves your fingers. Some points in that phase space will result in the coin landing heads up; color these points black. Other points in the phase space will result in the coin landing tails up; color these points white. If you examined the phase space under a microscope (metaphorically speaking) you would see an intricate pattern of black and white, with even a small movement in the phase space crossing many boundaries between a black region and a white region.

If you knew the initial conditions precisely enough, you would know whether the coin was in a white or black region of phase space, and you would then have a probability of either 1 or 0 for it coming up heads.

It’s more typical that we don’t have such precise measurements, and so we can only pin down the coin’s location in phase space to a region that contains many, many black subregions and many, many white subregions… effectively it’s just gray, and the shade of gray is your probability for heads given your measurement of the initial conditions.

So you see that the answer to “what is the probability of the coin landing heads up” depends on what information you have available.

Of course, in practice you typically don’t even have the lesser level of information assumed above—you don’t know enough about the coin, even in principle, to compute which points in phase space are black and which are white, or what proportion of the points are black versus white in the region corresponding to what you know about the initial conditions. Here’s where symmetry arguments then give you P(heads) = ¹⁄₂.

ksvanhorn 24 Jul 2018 4:17 UTC
22 points
on: Bayesian Probability is for things that are Space-like Separated from You
I’m not trying to be mean here, but this post is completely wrong at all levels. No, Bayesian probability is not just for things that are space-like. None of the theorems from which it derived even refer to time.
So, you know the things in your past, so there is no need for probability there.
This simply is not true. There would be no need of detectives or historical researchers if it were true.
If you partially observe a fact, then I want to say you can decompose that fact into the part that you observed and the part that you didn’t, and say that the part you observed is in your past, while the part you didn’t observe is space-like separated from you.
You can say it, but it’s not even approximately true. If someone flips a coin in front of me but covers it up just before it hits the table, I observe that a coin flip has occurred, but not whether it was heads or tails—and that second even is definitely within my past light-cone.
You may have cached that you should use Bayesian probability to deal with things you are uncertain about.
No, I cached nothing. I first spent a considerable amount of time understanding Cox’s Theorem in detail, which derives probability theory as the uniquely determined extension of classical propositional logic to a logic that handles uncertainty. There is some controversy about some of its assumptions, so I later proved and published my own theorem that arrives at the same conclusion (and more) using purely logical assumptions/requirements, all of the form, “our extended logic should retain this existing property of classical propositional logical.”
The problem is that the standard justifications of Bayesian probability are in a framework where the facts that you are uncertain about are not in any way affected by whether or not you believe them!
1) It’s not clear this is really true. It seems to me that any situation that is affected by an agent’s beliefs can be handled within Bayesian probability theory by modeling the agent.
2) So what?
Therefore, our reasons for liking Bayesian probability do not apply to our uncertainty about the things that are in our future!
This is a complete non sequitur. Even if I grant your premise, most things in my future are unaffected by my beliefs. The date on which the Sun will expand and engulf the Earth is in no way affected by any of my beliefs. Whether you will get luck with that woman at the bar next Friday is in no way affected by any of my beliefs. And so on,

ksvanhorn 10 Mar 2012 22:10 UTC
18 points
on: The Neglected Virtue of Scholarship
In my experience, Ph.D. dissertations can be a wonderful resource for getting an overview of a particular academic topic. This is because the typical—and expected—pattern for a dissertation is to first survey the existing literature before diving into one’s own research. This both shows that the doctoral candidate has done his/her homework, and, just as importantly, brings his/her committee members up to speed on the necessary background. For example, a lot of my early education in Bayesian methods came from reading the doctoral dissertations of Wray Buntine, David J. C. MacKay, and Radford Neal on applications of Bayesian methods to machine learning. Michael Kearns’ dissertation helped me learn about computational learning theory. A philosophy dissertation helped me learn about temporal logic.

Of course, this requires that you already have some background in some related discipline. My background was in computer science when I read the above-mentioned dissertations, along with a pretty good foundation in mathematics.

Sleeping Beauty Resolved?

ksvanhorn22 May 2018 14:13 UTC

16 points

77 comments10 min readLW link

ksvanhorn 20 Jan 2011 4:54 UTC
14 points
in reply to: kurige’s comment on: Moore’s Paradox
“I chose to believe in the existance of God—deliberately and conciously.”

I cannot conceive of how it is possible to deliberately and consciously choose to believe in something.

I grew up in a religious family. I served as a missionary for my church. I married a woman of the same religion. For most of my first 28 years I believed not only that there was a God but that he had established the church of which I and my family were all members.

But once I started examining my beliefs more closely, I realized that I was engaging in the most dishonest sort of special pleading in their favor. And then it was no longer possible to continue believing.

ksvanhorn 1 Mar 2012 20:44 UTC
12 points
on: Status and Changing your Mind
Abandoning your previous position can also be a way of saving face, in at least two ways:
- Being wrong is embarrassing, yes; but being wrong for a short period of time is less embarrassing than being wrong for an extended period of time. Best to stop the bleeding as quickly as possible.
- You signal that you are a reasonable person who does not let emotional attachment to a position cloud his judgment. If you’re dealing with someone of higher status, you show that your mistake doesn’t matter that much because you corrected yourself quickly. If you are dealing with someone who is lower status than you, you come off appearing magnanimous.

ksvanhorn 17 Jul 2011 15:27 UTC
12 points
in reply to: lessdazed’s comment on: The limits of introspection
It’s hard to maximize your expected utility if you don’t really know what your utility function is. And attempts at self-improvement may founder if you have an inaccurate picture of yourself.

ksvanhorn 10 Jun 2012 3:56 UTC
11 points
on: Conspiracy Theories as Agency Fictions

Learning about useful models helps people escape anthropomorphizing human society or the economy or government. The latter is particularly salient. I think most people slip up occasionally in assuming that say something like the United States government can be successfully modelled as a single agent to explain most of its “actions”.

As an interesting (to me, at least) aside, Gene Sharp’s research on nonviolent resistance indicates that successful nonviolent resistance invariably involves taking to heart this little idea—that governments are not single agents but systems of many agents pursuing their own ends—and exploiting it to the max.

ksvanhorn 13 Jan 2012 7:10 UTC
11 points
in reply to: adamisom’s comment on: Can the Chain Still Hold You?
even the Bayesian interpretation of probability theory has only been accepted for… well, I’m not sure, but I think only since World War II

Try half a century later. Until very recently—about twenty years ago—the Bayesian view of probability was very much a minority view, and it has only really picked up steam in the last 10 years. Several things happened around 20 years ago:
- Faster and cheaper computers became available. Bayesian methods tend to be computationally intensive, and this limited their use.
- Rule-based expert systems fizzled out and began to be replaced by Bayesian networks after practical algorithms for inference with BNs were developed.
- Awareness of Markov chain Monte Carlo methods (which can be used to sample from a Bayesian posterior distribution) spread to the statistics community, and the free BUGS software made it easy for non-experts to create and evaluate new Bayesian models.
These developments made it practical to apply Bayesian methods… and people started finding out how well they could work.

ksvanhorn 31 Dec 2011 1:30 UTC
11 points
on: Tell Your Rationalist Origin Story
Oddly enough, politics was the catalyst for me.

I grew up in a very religious, very conservative Mormon family. From my father I acquired the attitude that there are few things more shameful than dishonesty. From reading science fiction, particularly Asimov and Heinlein, and reading science books, I acquired the ideal of intellectual honesty. My father had very strong religious and political opinions that brooked no dissent. In attempting to formulate a consistent political philosophy of my own, I found my opinions diverging from his, but I lacked the courage to openly contradict him. After I had been away from home for several years, in my early twenties, I went through a period where I made a serious effort to root out any inconsistencies in my political philosophy and just honestly follow the consequences of my principles wherever they led. I ended up a libertarian anarchist.

I didn’t know it at the time, but that was the beginning of the end for my religious beliefs. Intellectual honesty had long been an ideal for me; now it was an important part of my self-image. I found that I could no longer ignore the special pleading I engaged in when it came to my religious beliefs. If I applied to my religious beliefs the same standards I used to evaluate non-religious claims, they started to look pretty shaky. But everyone in my family and everyone in my social circle was Mormon. I had spent a year and a half as a missionary for the Mormon church. My wife was a devout Mormon. I had just started my graduate studies at BYU, a university owned by the Mormon church.

And my father reserved his most vociferous condemnation for “apostates”.

The critical point came when I was 28, during an interview with my bishop for a temple recommended. One of the questions he asked was, “Do you believe in God the Father, and in his son Jesus Christ, and in the Holy Ghost?” I realized that, in fact, I did not, and my upbringing did not allow me to lie about it. I went home without a temple recommend… and then had to explain to my wife why.

Everything else since then has just been filling in the details.

ksvanhorn 25 Jul 2018 17:23 UTC
10 points
in reply to: dxu’s comment on: Bayesian Probability is for things that are Space-like Separated from You
This is the first thing I’ve read from Scott Garrabant, so “otherwise reputable” doesn’t apply here. And I have frequently seen things written on LessWrong that display pretty significant misunderstandings of the philosophical basis of Bayesian probability, so that gives me a high prior to expect more of them.

ksvanhorn 21 Jan 2012 17:34 UTC
10 points
on: Some potential dangers of rationality training

...many people’s lives aren’t much improved by rationality training.

The outlook for rationality training may be brighter than you think. I’ve been looking at a number of materials aimed at businessmen who want greater success, and a lot of the techniques described and used by successful businessmen bear a strong resemblance to the sort of things you read on Less Wrong. For example, consider this passage in Peter Drucker’s book The Effective Executive, describing Alfred P. Sloan, a legendary CEO of GM:

Alfred P. Sloan is reported to have said at a meeting of one of his top committees: “Gentlemen, I take it we are all in complete agreement on the decision here.” Everyone around the table nodded assent. “Then,” continued Mr. Sloan, “I propose we postpone further discussion of this matter until our next meeting to give ourselves time to develop disagreement and perhaps gain some understanding of what the decision is all about.”

Sloan was anything but an “intuitive” decision-maker. He always emphasized the need to test opinions against facts and the need to make absolutely sure that one did not start out with the conclusion and then look for the facts that would support it. But he knew that the right decision demands adequate disagreement.

So here we have talk of the need to consider multiple alternatives, the importance of not just getting one side of a story, the need to test hypotheses against reality, and the importance of not starting with the bottom line… all familiar topics in our community.

Sleeping Beauty Resolved (?) Pt. 2: Identity and Betting

ksvanhorn2 Jun 2018 2:43 UTC

9 points

50 comments6 min readLW link

ksvanhorn 26 Dec 2011 1:48 UTC
9 points
on: [Link] Correlation Graphs Reveal Shocking Information
I don’t in any way intend this as a criticism, but Mark Twain’s complaint about bringing up the weather in conversation seems to apply here: everyone talks about it, but nobody ever does anything about it.

The problem with just saying “correlation is not causation” is that it doesn’t help you figure out what information you can get from an observational study. This is becoming an important issue in my work. We do advanced marketing research for (mostly) large companies. For many years the company’s emphasis was on choice-based conjoint studies, which give you experimental data. (You ask people to repeatedly choose among various sets of hypothetical products, and then analyze the results to figure out what they value.) Now we’re moving more into marketing mix models, which involve purely observational data. Knowing exactly what one can and cannot legitimately infer from observational data, under what assumptions, is a very important practical question for us and our clients.

That’s why I am currently spending a lot of my spare time studying Judea Pearl’s book Causality. I would highly recommend this book to anyone who, as Yvain suggests, is more interested in solving problems than looking smart.

ksvanhorn 25 Dec 2011 21:25 UTC
9 points
in reply to: Vaniver’s comment on: Applied Rationality Practice
A note on the recommendation to study physics: physicists can do anything. Seriously. You see physicists making contributions to computer science, biology, statistics, computational finance, etc. You rarely see non-physicists making significant contributions to physics. I think the reason for this is that physicists learn lots of very useful mathematics that can be applied in a wide variety of contexts… and unlike someone who earns a mathematics degree, their emphasis is on applying math to solving problems, rather than on math for its own sake. The other factor may be that physicists are trained to think in terms of fundamental principles; they expect to find some hidden underlying pattern that will bring order out of the chaos.