This article is an attempt to summarize basic material, and thus probably won't have anything new for the hard core posting crowd. It'd be interesting to know whether you think there's anything essential I missed, though.

You've probably seen the word 'Bayesian' used a lot on this site, but may be a bit uncertain of what exactly we mean by that. You may have read the intuitive explanation, but that only seems to explain a certain math formula. There's a wiki entry about "Bayesian", but that doesn't help much. And the LW usage seems different from just the "Bayesian and frequentist statistics" thing, too. As far as I can tell, there's no article explicitly defining what's meant by Bayesianism. The core ideas are sprinkled across a large amount of posts, 'Bayesian' has its own tag, but there's not a single post that explicitly comes out to make the connections and say "this is Bayesianism". So let me try to offer my definition, which boils Bayesianism down to three core tenets.

We'll start with a brief example, illustrating Bayes' theorem. Suppose you are a doctor, and a patient comes to you, complaining about a headache. Further suppose that there are two reasons for why people get headaches: they might have a brain tumor, or they might have a cold. A brain tumor always causes a headache, but exceedingly few people have a brain tumor. In contrast, a headache is rarely a symptom for cold, but most people manage to catch a cold every single year. Given no other information, do you think it more likely that the headache is caused by a tumor, or by a cold?

If you thought a cold was more likely, well, that was the answer I was after. Even if a brain tumor caused a headache every time, and a cold caused a headache only one per cent of the time (say), having a cold is so much more common that it's going to cause a lot more headaches than brain tumors do. Bayes' theorem, basically, says that if cause A might be the reason for symptom X, then we have to take into account both the probability that A caused X (found, roughly, by multiplying the frequency of A with the chance that A causes X) and the probability that anything else caused X. (For a thorough mathematical treatment of Bayes' theorem, see Eliezer's Intuitive Explanation.)

Followup to:  The Outside View's Domain, Conversation Halters
Reply to:  Reference class of the unclassreferenceable

In "conversation halters", I pointed out a number of arguments which are particularly pernicious, not just because of their inherent flaws, but because they attempt to chop off further debate - an "argument stops here!" traffic sign, with some implicit penalty (at least in the mind of the speaker) for trying to continue further.

This is not the right traffic signal to send, unless the state of knowledge is such as to make an actual halt a good idea.  Maybe if you've got a replicable, replicated series of experiments that squarely target the issue and settle it with strong significance and large effect sizes (or great power and null effects), you could say, "Now we know."  Or if the other is blatantly privileging the hypothesis - starting with something improbable, and offering no positive evidence to believe it - then it may be time to throw up hands and walk away.  (Privileging the hypothesis is the state people tend to be driven to, when they start with a bad idea and then witness the defeat of all the positive arguments they thought they had.)  Or you could simply run out of time, but then you just say, "I'm out of time", not "here the gathering of arguments should end."

But there's also another justification for ending argument-gathering that has recently seen some advocacy on Less Wrong.

An experimental group of subjects were asked to describe highly specific plans for their Christmas shopping:  Where, when, and how.  On average, this group expected to finish shopping more than a week before Christmas.  Another group was simply asked when they expected to finish their Christmas shopping, with an average response of 4 days.  Both groups finished an average of 3 days before Christmas.  Similarly, Japanese students who expected to finish their essays 10 days before deadline, actually finished 1 day before deadline; and when asked when they had previously completed similar tasks, replied, "1 day before deadline."  (See this post.)

Those and similar experiments seem to show us a class of cases where you can do better by asking a certain specific question and then halting:  Namely, the students could have produced better estimates by asking themselves "When did I finish last time?" and then ceasing to consider further arguments, without trying to take into account the specifics of where, when, and how they expected to do better than last time.

From this we learn, allegedly, that "the 'outside view' is better than the 'inside view'"; from which it follows that when you're faced with a difficult problem, you should find a reference class of similar cases, use that as your estimate, and deliberately not take into account any arguments about specifics.  But this generalization, I fear, is somewhat more questionable...

Related to:  Logical Rudeness, Semantic Stopsigns

While working on my book, I found in passing that I'd developed a list of what I started out calling "stonewalls", but have since decided to refer to as "conversation halters".  These tactics of argument are distinguished by their being attempts to cut off the flow of debate - which is rarely the wisest way to think, and should certainly rate an alarm bell.

Here's my assembled list, on which I shall expand shortly:

• Appeal to permanent unknowability;
• Appeal to humility;
• Appeal to egalitarianism;
• Appeal to common guilt;
• Appeal to inner privacy;
• Appeal to personal freedom;
• Appeal to arbitrariness;
• Appeal to inescapable assumptions.
• Appeal to unquestionable authority;
• Appeal to absolute certainty.

Now all of these might seem like dodgy moves, some dodgier than others.  But they become dodgier still when you take a step back, feel the flow of debate, observe the cognitive traffic signals, and view these as attempts to cut off the flow of further debate.

Daniel Dennett has advanced the opinion that the evolutionary purpose of the cuteness response in humans is to make us respond positively to babies.  This does seem plausible.  Babies are pretty cute, after all.  It's a tempting explanation.

Here is one of the cutest baby pictures I found on a Google search.

And this is a bunny.

Correct me if I'm wrong, but the bunny is about 75,119 times cuter than the baby.

Now, bunnies are not evolutionarily important for humans to like and want to nurture.  In fact, bunnies are edible.  By rights, my evolutionary response to the bunny should be "mmm, needs a sprig of rosemary and thirty minutes on a spit".  But instead, that bunny - and not the baby or any other baby I've seen - strikes the epicenter of my cuteness response, and being more baby-like along any dimension would not improve the bunny.  It would not look better bald.  It would not be improved with little round humanlike ears.  It would not be more precious with thumbs, easier to love if it had no tail, more adorable if it were enlarged to weigh about seven pounds.

If "awwww" is a response designed to make me love human babies and everything else that makes me go "awwww" is a mere side effect of that engineered reaction, it is drastically misaimed.  Other responses for which we have similar evolutionary psychology explanations don't seem badly targeted in this way.  If they miss their supposed objects at all, at least it's not in most people.  (Furries, for instance, exist, but they're not a common variation on human sexual interest - the most generally applicable superstimuli for sexiness look like at-least-superficially healthy, mature humans with prominent human sexual characteristics.)  We've invested enough energy into transforming our food landscape that we can happily eat virtual poison, but that's a departure from the ancestral environment - bunnies?  All natural, every whisker.¹

Followup to:  Logical Rudeness

"Modern man is so committed to empirical knowledge, that he sets the standard for evidence higher than either side in his disputes can attain, thus suffering his disputes to be settled by philosophical arguments as to which party must be crushed under the burden of proof."
        -- Alan Crowe

There's a story - in accordance with Poe's Law, I have no idea whether it's a joke or it actually happened - about a creationist who was trying to claim a "gap" in the fossil record, two species without an intermediate fossil having been discovered.  When an intermediate species was discovered, the creationist responded, "Aha!  Now there are two gaps."

Since I'm not a professional evolutionary biologist, I couldn't begin to rattle off all the ways that we know evolution is true; true facts tend to leave traces of themselves behind, and evolution is the hugest fact in all of biology.  My specialty is the cognitive sciences, so I can tell you of my own knowledge that the human brain looks just like we'd expect it to look if it had evolved, and not at all like you'd think it would look if it'd been intelligently designed.  And I'm not really going to say much more on that subject.  As I once said to someone who questioned whether humans were really related to apes:  "That question might have made sense when Darwin first came up with the hypothesis, but this is the twenty-first century.  We can read the genes.  Human beings and chimpanzees have 95% shared genetic material.  It's over."

Well, it's over, unless you're crazy like a human (ironically, more evidence that the human brain was fashioned by a sloppy and alien god).  If you're crazy like a human, you will engage in motivated cognition; and instead of focusing on the unthinkably huge heaps of evidence in favor of evolution, the innumerable signs by which the fact of evolution has left its heavy footprints on all of reality, the uncounted observations that discriminate between the world we'd expect to see if intelligent design ruled and the world we'd expect to see if evolution were true...

...instead you search your mind, and you pick out one form of proof that you think evolutionary biologists can't provide; and you demand, you insist upon that one form of proof; and when it is not provided, you take that as a refutation.

You say, "Have you ever seen an ape species evolving into a human species?"  You insist on videotapes - on that particular proof.

And that particular proof is one we couldn't possibly be expected to have on hand; it's a form of evidence we couldn't possibly be expected to be able to provide, even given that evolution is true.

Yet it follows illogically that if a video tape would provide definite proof, then, likewise, the absence of a videotape must constitute definite disproof.  Or perhaps just render all other arguments void and turn the issue into a mere matter of personal opinion, with no one's opinion being better than anyone else's.

A few of us got together in the pub after the friendly AI meet and agreed we should have a meetup for those of us familiar with lesswrong/bostrom etc. This is a post for discussion of when/where.

The Large Hadron Collider was shut down yesterday by a hamster in a tutu, weary scientists announced.

The Large Hadron Collider is the successor to the earlier Superconducting Super Collider, which was shut down by the US House of Representatives in 1993 after 14 miles of tunnel had been constructed at a cost of $2 billion.  Since its inception, the Large Hadron Collider has been plagued by construction delays, dead technicians, broken magnet supports, electrical faults, helium containment failures, vacuum leaks, birds with baguettes, terrorists, ninjas, pirates, supervillains, hurricanes, asteroids, cosmic energy storms, and a runaway train.  On one occasion it was discovered that the entire 17-mile circular tunnel had been built upside-down due to a sign error in the calculations, and the whole facility had to be carefully flipped by a giant spatula.

People debate all the time about how strictly children should be disciplined. Obviously, this is a worthwhile debate to have, as there must be some optimal amount of discipline that is greater than zero. The debate's nominal focus is usually on what's best for the child, with even the advocates for greater strictness arguing that it's "for their own good." It might also touch on what's good for other family members or for society at large. What I think is missing from the usual debate is that it assumes nothing but honorable motives on the part of the arguers. That is, it assumes that the arguments in favor of greater strictness are completely untainted by any element of authoritarianism or cruelty. But people are sometimes authoritarian and cruel! Just for fun! And the only people who you can be consistently cruel to without them slugging you, shunning you, suing you, or calling the police on you are your children. This is a reason for more than the usual amount of skepticism of arguments that say that strict parenting is necessary. If there were no such thing as cruelty in the world, people would still argue about the optimal level of strictness, and sometimes the more strict position would be the correct one, and parents would chose the optimal level of strictness on the basis of these arguments. But what we actually have is a world with lots and lots of cruelty lurking just under the surface, which cannot help but show up in the form of pro-strictness arguments in parenting debates. This should cause us to place less weight on pro-strictness arguments than we otherwise would.*  Note that this is basically the same idea as Bertrand Russell's argument against the idea of sin: its true function is to allow people to exercise their natural cruelty while at the same time maintaining their opinion of themselves as moral.

I've been trying for a while to make sense of the various alternate decision theories discussed here at LW, and have kept quiet until I thought I understood something well enough to make a clear contribution.  Here goes.

You simply cannot reason about what to do by referring to what program you run, and considering the other instances of that program, for the simple reason that: there is no unique program that corresponds to any physical object.

Yes, you can think of many physical objects O as running a program P on data D, but there are many many ways to decompose an object into program and data, as in O = <P,D>.  At one extreme you can think of every physical object as running exactly the same program, i.e., the laws of physics, with its data being its particular arrangements of particles and fields.  At the other extreme, one can think of each distinct physical state as a distinct program, with an empty unused data structure.  Inbetween there are an astronomical range of other ways to break you into your program P and your data D.

Eliezer's descriptions of his "Timeless Decision Theory", however refer often to "the computation" as distinguished from "its input" in this "instantiation" as if there was some unique way to divide a physical state into these two components.  For example:

The one-sentence version is:  Choose as though controlling the logical output of the abstract computation you implement, including the output of all other instantiations and simulations of that computation.

The three-sentence version is:  Factor your uncertainty over (impossible) possible worlds into a causal graph that includes nodes corresponding to the unknown outputs of known computations; condition on the known initial conditions of your decision computation to screen off factors influencing the decision-setup; compute the counterfactuals in your expected utility formula by surgery on the node representing the logical output of that computation.

And also:

Timeless decision theory, in which the (Godelian diagonal) expected utility formula is written as follows:  Argmax[A in Actions] in Sum[O in Outcomes](Utility(O)*P(this computation yields A []-> O|rest of universe))  ... which is why TDT one-boxes on Newcomb's Problem - both your current self's physical act, and Omega's physical act in the past, are logical-causal descendants of the computation, and are recalculated accordingly inside the counterfactual. ...  Timeless decision theory can state very definitely how it treats the various facts, within the interior of its expected utility calculation.  It does not update any physical or logical parent of the logical output - rather, it conditions on the initial state of the computation, in order to screen off outside influences; then no further inferences about them are made.

These summaries give the strong impression that one cannot use this decision theory to figure out what to decide until one has first decomposed one's physical state into one's "computation" as distinguished from one's "initial state" and its followup data structures eventually leading to an "output."  And since there are many many ways to make this decomposition, there can be many many decisions recommended by this decision theory. 

The advice to "choose as though controlling the logical output of the abstract computation you implement" might have you choose as if you controlled the actions of all physical objects, if you viewed the laws of physics as your program, or choose as if you only controlled the actions of the particular physical state that you are, if every distinct physical state is a different program.