NatashaRostova

• NatashaRostova 23 Nov 2016 22:36 UTC

  2 points

  in reply to: Sophivorus’s comment on: Can any­one be ra­tio­nal and not ve­gan?

  My utility from eating meat is b(X=1) My utility from the world not being destroyed is Y So...

  Utility = X + E[Y|X=1] vs. Utility = E[Y| X=0]

  Obviously you’re able to disagree with my model. But in my self-consistent model of the world the benefit I get from eating meat is greater than the utility I achieve from lowering the likelihood the world is destroyed (in any given timeframe) by not eating meat.

  I strongly suspect you will have a hard time convincing me I’m wrong without appealing to moral imperatives. Obviously there isn’t some mathematically elegant and clearly distinct line between rationality and moral imperative, but hey, if there was Less Wrong probably wouldn’t exist.

• NatashaRostova 24 Nov 2016 23:05 UTC

  0 points

  on: Mis­matched Vocabularies

  #3 Can be decomposed into many parts. Most of them contingent on your relationship with the person, and your use of terminology.

  As a sort of aside, being an effective communicator can depend on your understanding the other persons frame of reference, and working to not use terms they are unlikely to know. This often depends on specific social dynamics, but you can frame it like “In my field there is this concept XYZ, I’m sure you’ve heard of it, and blah blah.” This way you give them the benefit of the doubt, but also give them a way to not feel bad as you note it’s in your field. I know this might feel strange to do when you yourself never take stuff personally—to that I can only offer sympathy.

  ...On the other hand, most complicated terminology or jargon often references a semi-complex phenomena. It’s often not fair, at all, to expect others to know or follow it. Sometimes the person speaking uses it to give their argument an authoritative veneer, because the other person doesn’t know or can’t respond. If people get that impression they sometimes get angry.

  Anyway, I’d encourage you to maybe be more empathetic towards other people in these instances. People often are ‘crazy,’ but it’s best to assume they aren’t until you’ve really, really, tried to see their model of the world.

• NatashaRostova 24 Nov 2016 23:20 UTC

  0 points

  in reply to: Daniel_Burfoot’s comment on: Open thread, Nov. 21 - Nov. 27 − 2016

  I am not a ML Research Scientist—but have studied and used it, and find it very interesting.

  As far as I understand, deep learning is able to discover and fit any nonlinear dynamics in a set of data, which then of course has to be trained/​regularized/​cross-validated to prune away over-fitting. If we accept the view that reality is just a huge set of nonlinear equations and information, and NN/​DL can discover these at any level of granularity, then it is a reasonable prediction that they are well posed to be the best.

  Also I’m not confident in this next point, but would love some additional feedback. Read this part with skepticism: As I understand effective DL works so well because it combines filtering and tractability within the model structure, and with variation on layering/​neurons/​optimization techniques, it opens up a greater set of potential models than many other model classes. This makes the fact that it works so well with GPU not a lucky accident, but rather an intrinsic feature of the mathematical structure of the model. Perhaps that’s why we evolved to use NN type structure in our brain—due to its tractability and parallel information processing abilities?

  To use an example from my own research, in Financial Econometric asset pricing we often use this tool called a Kalman Filter to filter out states of the world from sets of stochastic PDEs. Optimizing those models, when they have more than ~10 parameters, is such a hassle. It requires lots of optimization black-magic, which is a quasi-scientific method where over months you run different optimization algorithms on the whole model, then single parameters, then the whole model, and if some parameter looks ‘weird’ you manually change it to what you ‘think’ it should be.

  Basically a neural network could learn the dynamics here (without revealing them to us), and provide a potentially equal forecast (haven’t tested this). It could also do this much faster than our optimization method (I think). The forecast would be less useful to a human analyst, because without the model dynamics made explicit, it is much harder to run simulations and study specific parameters. But it’s within the realm of reason to predict that a computer, were it self aware, would be able to understand the way the parameters work in the NN/​DL model itself.

  For this reason I think the natural structure of the model makes it somewhat true that it is #2, but that this makes it a special case of #1 (as MrMind pointed out before me).

  Again, I’m not a ML Research Scientist—so if I’ve totally messed something up I’d love to know what and why.

Nas­sim Taleb on Elec­tion Forecasting

• NatashaRostova 27 Nov 2016 3:29 UTC

  0 points

  in reply to: Daniel_Burfoot’s comment on: Open thread, Nov. 21 - Nov. 27 − 2016

  To address your third point first, I’m sure you are right. I have only played around with simple NNs, and shouldn’t have spoken freely on how it would be easy to estimate a more complex one, when I don’t know much about it.

  As a follow up question to your second point: The Kalman filter is a very aesthetically pleasing model, I agree. Something I wonder, but have no idea on, is whether there are mathematical concepts similar to the Kalman filter (in terms of aesthetics and usefulness) that are entirely outside of the understanding of the human brain. So, hypothetically, if we engineered humans with IQ 200+ (or whatever), they would uncover things like the Kalman Filter that normal humans couldn’t grasp.

  If that’s true, does it stand to reason that we could still use those models with a sufficiently well optimized/​built DNN? We would just never understand what’s going on inside the network?

  I often think of self-driving cars as learning the dynamic interactions of a set of nonlinear equations that are beyond the scope of a human to ever derive.

  I’ll note I realize some of my questions might be too vague or pseudo-philosophical to be answered.

  PS: I did a little internet sleuthing and have read the first ~12 pages of your book so far, which is very interesting and similar to how I think of the world (yours is much more well developed). I am also incredibly interested in empirical philosci and read/​write/​think about it a ton.

• NatashaRostova 27 Nov 2016 22:35 UTC

  2 points

  in reply to: knb’s comment on: Nas­sim Taleb on Elec­tion Forecasting

  I want to be careful in how I talk about Adams. He definitely didn’t follow the guidelines for methodological forecasting, such as assigning clear numerical predictions and tracking a Brier (or any chosen) scoring method.

  As a result I see two main groups of thought on Adams: The first is forecasting oracle. The second is total charlatan (as far as I can tell this is the Rationalist viewpoint, I know SSC took this view).

  I think the rationalist viewpoint is close to right. If we include the set of all semi-famous people who did/​could speculate on an election (including Adams), and then imagine (we don’t have the data) that we tracked all their predictions, with the knowledge that after the fact we would forget everyone who was wrong, Adams doesn’t seem significantly correct.

  But if Adams (or an abstracted idea of Adams argument) were correct, It would be because unlike current polling methods it allows for really high-dimensional data to be embedded into the forecast. As of now humans seem to be much better at getting a ‘feel’ for a movement than computers, because it requires using vast unrelated and unstructured data, which we specifically evolved to do* (I know we don’t have great experiments to determine what we did/​didn’t specifically evolve for, so ignore this point if you want).

  So, to that extent, current purely model-based election forecasts are at risk of having a severe form of omitted variable bias.

  As an example, while polls are a little stable, Marine Le Pen is currently at a huge disadvantage: “According to a BVA poll carried out between Oct. 14 and Oct. 19, Le Pen would win between 25 percent and 29 percent of the vote in next April’s first round. If she faces Bordeaux mayor Alain Juppe—the favorite to win the Republicans primary—she’d lose the May 7 run-off by more than 30 percentage points. If it’s former President Nicolas Sarkozy, the margin would be 12 points.”*

  And yet PredictIt.org has her at ~40%. There is strong prior information from Brexit/​​Trump that seems important, but is absent in polls. It’s almost as if we are predicting how people will change their mind when exposed to a ‘treatment effect’ of rightwing nationalism.

  *http://​​www.bloomberg.com/​​news/​​articles/​​2016-11-16/​​french-pollsters-spooked-by-trump-but-still-don-t-see-le-pen-win

  So then to tie this back to the original post, if you have stronger prior information, such as a strong reason to believe races will be 50-50, non-uniform priors, or that omitted variable bias exists, it would make sense to impose a structure on time-variation of the poll. I think this set of reasons is why it feels wrong to us when we see predictions varying so much far off from an election.

• NatashaRostova 28 Nov 2016 1:19 UTC

  3 points

  on: Stand-up com­edy as a way to im­prove ra­tio­nal­ity skills

  I think you’re definitely on to something interesting. A certain type of clever humor is associated with exposing strange patterns that seem nonsensical or aren’t examined, which for whatever strange reason we find funny. Wordplay is part of this as well (Scott Alexander has a whole book based on humorous wordplay, among other topics).

  I would strongly suspect humor doesn’t make you more rational/​smarter/​scientific, but is a fun grounds for playing around with strange ideas and patterns.

  Having said that, I can think of worst ways to spend a Friday night :)

• NatashaRostova 28 Nov 2016 2:34 UTC

  0 points

  on: On call­ing in­con­ceiv­able what you’ve already con­ceived.

  Word thinking is better at getting at weird things. Weird or not, it’s there.

  Another way I think of it is, imagine a function f(x), that is unique for all values, which returns the same value for x=3 and x=10.

  That’s sort of like… Okay… So there is some function with two properties that imply the other cannot exist. What does it mean?!

  I think lots of these philosophical problems were very reasonable to discuss when they were first brought up. I can’t prove this, since it’s based on lots of general concepts I’ve found in old books or academic papers, but those ideas seem related to the same idea of the human brain as mystic, and language an expression of that mysticism. Rather than an evolved mechanism to filter patterns from reality and communicate them to one another. In that anti-mystical view, which views words and math as the same subset of pattern classification and information, poorly defined thought experiments are as interesting as “what if 2+2=5”.

  Also, I want to speculate on an idea I have with near zero proof that I half stole from/​half conceived from reading “Godel, Escher, Bach:” We know computers cannot process undefined statements. When humans try to process undefined statements we don’t crash. Instead we conceive something that seems like it could be meaningful, even if it’s inconceivable. I can ‘conceive’ of 2+2=5, or it feels like I can, but maybe that’s just the error-catching subroutine of my brain. We then sort of interpret this inconsistency as meaningful and call it a ‘thought experiment’—when the inconsistency is just our brain’s way of saying “does not compute.”

• NatashaRostova 28 Nov 2016 20:08 UTC

  4 points

  in reply to: ingive’s comment on: Open thread, Nov. 28 - Dec. 04, 2016

  You can’t just classify things as stupid because you think they are stupid, and what you think is stupid is true because you think you’re rational.

  The idea that ‘not being vegan’ or ‘smoking’ are stupid is silly.

• NatashaRostova 28 Nov 2016 21:52 UTC

  0 points

  on: On the im­por­tance of Less Wrong, or an­other sin­gle con­ver­sa­tional locus

  I’ve known about Less Wrong for about two full years. A few weeks ago I started coming here regularly. A week ago I made an account—right before this post and others like it.

  My own poetic feeling is there is a change in the winds, and the demand for a good community is growing. SSC has no real community. Facebook is falling apart with fake news and awful political memes. People are losing control of their emotions w.r.t politics. And calm scientific rationalist approaches are falling apart.

  I deactivated my FB, made an account here, and have done my best, despite not (yet) being anyone well known, to keep it going.

  I think there is a change that is drawing people here again. But it needs to be encouraged. The rules for rationalist discussion 5 years ago aren’t the ones of today. Get rid of the stupid no politics rule, and just enforce good scientific discussion. Scott Alexander’s Trump post has received 200k+ views. The demand for a well written post like that was incredible, and the original ‘rationalist’ heart of the internet wouldn’t let it be written. Is that not silly?

• NatashaRostova 28 Nov 2016 22:02 UTC

  0 points

  in reply to: Alexandros’s comment on: On the im­por­tance of Less Wrong, or an­other sin­gle con­ver­sa­tional locus

  Further, I am fairly certain that LW as a community blog is bound to fail. Strong writers enjoy their independence.

  That’s true. LW isn’t bringing back yvain/​Scott or other similar figures. However, it is a cool training ground/​incubator for aspiring writers. As of now I’m a ‘no one.’ I’d like to try to see if I can become ‘some one.’ SSC comments don’t foster this. LW is a cool place to try, it’s not like anyone is currently reading my own site/​blog.

• NatashaRostova 29 Nov 2016 6:30 UTC

  8 points

  on: Ex­pert Pre­dic­tion Of Experiments

  I guess really the only conclusion you can draw from all of this is not to put any important decisions in the hands of people from top business schools.

  I used to strongly agree with this point, but now I worry that dismissal of MBA skills has more to do with a failure of our ability to scientifically explain their mechanism, rather than MBAs actually generally sucking.

  The problem space of probabilistic predictions is very distinct. In this way it lends itself to a combination of applied math and heuristics, which Tetlock lays out in Supreforecasting and Expert Political Judgement.

  The MBA problem space is much more chaotic. I was on a team working on new initiatives at [Mega Tech Company], and the problem space was sparse, not much or any historical data, and no way for me to use my scientific or programming skills to come to clear conclusions. The challenge was that it required more of a simulation style of thinking, which is how I now think of MBAs.

  The MBAs would discuss random high-dimensional, unstructured, and piece-meal data. They would then use their incredibly powerful human brains to simulate whole sets of actions, how they would interact with the company, the customers, and the goals. and filter through an estimated probability space to choose what they thought is the right business choice. Sure, some of them sucked. But it was humbling to see the good ones.

  The reality is good MBAs (and people who self-select into MBAs) aren’t the type who care about careful analysis for one-shot predictions. They are, as far as I can tell, people with high IQs, a solid set of training data on Real Business Cases, and a ‘gut feeling’ capable of generating a probability distribution over high-dimensional probability spaces.

  You don’t want your president saying “I estimate there is an 80% chance of nuclear war.” You want them thinking “What is the probability space given this overwhelming amount of information, and how do I steer my country through the right conditions in this n-dimensional choice set.” (which I base on John Cochrane’s response to Tetlock in the Cato journal “What’s Wrong with Expert Predictions?” (https://​​www.cato-unbound.org/​​2011/​​07/​​15/​​john-h-cochrane/​​defense-hedgehogs) )*

• NatashaRostova 29 Nov 2016 18:24 UTC

  2 points

  in reply to: Douglas_Knight’s comment on: Nas­sim Taleb on Elec­tion Forecasting

  Thanks for the insight on the difference between Predictit/​Betfair—I wasn’t aware of this liquidity difference. Although so long as there is a reasonable amount of liquidity on Predict it, it’s very strange the two are not in equilibrium. Do you know if there are any open theories as to why this is?

  One thing I notice is a lot of commenters on PredictIt are alt-right/​NRx. It seems unlikely, but I wonder if different ideological priors are pushing different prediction markets away from a common equilibrium probability.

• NatashaRostova 30 Nov 2016 17:47 UTC

  0 points

  in reply to: Dagon’s comment on: Open thread, Nov. 28 - Dec. 04, 2016

  It doesn’t really matter what either of us think. If someone eats too much meat, and wishes they could stop, but can’t, then for a certain function we can claim it’s irrational in their achievement of that goal. If I eat a fair amount of meat because I work out, because it helps me get my weight-lifting goals, it’s rational for my objective.

  What’s your objective? Well, my main point is really just that we can’t abstract these sorts of things, they are empirical. “Is X irrational (implied: for all people under all conditions)?” is about as meaningful as “Does this chair really exist?”

• NatashaRostova 1 Dec 2016 17:36 UTC

  3 points

  on: Hate Crimes: A Fact Post

  To get that experience, you only need a few dozen actual recorded lynchings per year. The indirect impact of living under threat of violence far exceeds the literal death count.

  I think you demonstrated this point very clearly, which is something I have struggled with. It always frustrates me when people point out that “More people die from [banal, boring, cause] than [hot-button issue]” as though it were revealing some implication that we are ‘irrationally’ fearing a certain issue. (It is still a useful point though to remind people of how seemingly big numbers of deaths need to be compared to a huge population) The reality is that these numbers are the observed and measured worst outcomes of a system that spans a large spectrum of negative outcomes, and is hard to measure.

• NatashaRostova 1 Dec 2016 22:28 UTC

  1 point

  in reply to: Lumifer’s comment on: Hate Crimes: A Fact Post

  This point, and the one you make below, are fair points. I think the argument “we are overreacting to terrorism” seems like common sense or a self-evident argument.

  The reason I think it’s hard to measure is that suffering or reactions to perceived fear is still real suffering and real reactions. On the other hand, while lots of kids drown in pools or lose limbs from lawnmowers, no one suffers or reacts. (although one thing to consider is potential damage of terrorism is unbounded, whereas future changes in lawnmower deaths probably won’t spike).

  For the LW demographic, we probably don’t actually feel fear from terrorism. On the other hand, it makes me more angry than pool deaths or needless deaths due to no organ transplant free market. I often wonder: Can I override my base biological instinct that hates outsiders who attack? I can try to reason my way out, but that visceral anger response is hard-coded into my brain. I can, sort of. I still feel the hot-blooded anger, but I recognize it’s an evolutionary miscalibration with the modern age.

  At least I, we, have a level of reflection that comes from (typically) higher levels of education, intelligence, and often relative affluence.

  That’s why I consider it hard to measure. If I could prevent 800 deaths from terrorism, or 800 deaths from pool drownings, I would definitely prevent the first, because I know the negative effects would be much worse.

  I have a friend who has been too scared to leave her house for weeks since Trump got elected. She is suffering, even though it seems irrational to me, it’s rational within her own brain.

  So that’s what bugs me: We still have to measure her suffering, even if it is (probably) nonsensical. But if we try to optimize society with respect to our (argued) irrational suffering, we might end up with bad policy outcomes!

• NatashaRostova 2 Dec 2016 0:28 UTC

  1 point

  in reply to: Sable’s comment on: Which ar­eas of ra­tio­nal­ity are un­der­ex­plored? - Dis­cus­sion Thread

  Join the SlateStarCodex IRC :)

• NatashaRostova 2 Dec 2016 21:47 UTC

  0 points

  on: Ques­tion about metaethics

  It’s a good sign that you think it’s going over your head. Most of the time when people think “metaethics” (or metaphysics, or even normal ethics) is going over their head, it’s a sign that it’s actually nonsense. It’s always awkward trying to disprove nonsense, as dope Australian philosopher David Stove argued (http://​​web.maths.unsw.edu.au/​​~jim/​​wrongthoughts.html).

  EY and a few others have the ability to gaze into the technological machinations of a distant future involving artificial minds and their interactions with the coolest theoretical physics engineering feats. It’s clear to all smart people this is an inevitable conclusion, right? Fine.

  Okay going back to ethics, the point of the post (sorry, metaethics). Moral relativism is nothing more than a concept we hold in our minds. Insofar as it classifies different human beliefs on the world, and predicts their actions, it’s a useful term. It has no particularly profound meaning otherwise. It’s nothing more than a personal belief on how others should behave. You can’t test moral relativism, it has no fundamental property. The closest you can get to testing it is, as I just noted, asking how it predicts different human behaviors.

  Again, you tried to break this down which is understandable. But it’s not possible to refute or breakdown absolute nonsense. Some paperclip maximize doesn’t have values? So it won’t respond to some sort of ‘argument’ (which is a anthromorphic nonsensical set of information for the paperclip maximizer). And somehow this now connects to an argument that some other species will have some value, but it might be a bad one?

  Please let me know if you think I’m missing something, or some context from previous stuff he’s written that changed my interpretation of the writing above.

• NatashaRostova 2 Dec 2016 22:37 UTC

  5 points

  0

  in reply to: Vladimir_Golovin’s comment on: Mak­ing in­ten­tions con­crete—Trig­ger-Ac­tion Planning

  The zombie one is good, and something I will try. Sometimes when it’s 10pm and I’m pretty tired, I’ll try to work on some personal high-functioning project/​code, and get 3 minutes of work done by 11pm. On the other hand, if I used that time to read a book I could probably get more out o fit.

• NatashaRostova 3 Dec 2016 0:26 UTC

  2 points

  in reply to: Vladimir_Nesov’s comment on: Fact Posts: How and Why

  Check out Yvain’s sequence on Game Theory. I’ve actually studied game theory at a grad level, and had nothing to learn from what he wrote. But he opened it up in a fun/​interesting/​well-written way, which was specifically written for this audience, and addressed relevant interests here more than a textbook.

  It’s challenging to imagine a sequence on introductory chemistry that would have the same appeal. Having said that, I’m sure a sufficiently educated/​talented writer could do one on intro chem.