Posts

Moral realism and AI alignment 2018-09-03T18:46:44.266Z · score: 13 (4 votes)
The law of effect, randomization and Newcomb’s problem 2018-02-15T15:31:56.033Z · score: 18 (4 votes)
Naturalized induction – a challenge for evidential and causal decision theory 2017-09-22T08:15:09.999Z · score: 11 (9 votes)
A survey of polls on Newcomb’s problem 2017-09-20T16:50:08.802Z · score: 2 (2 votes)
Invitation to comment on a draft on multiverse-wide cooperation via alternatives to causal decision theory (FDT/UDT/EDT/...) 2017-05-29T08:34:59.311Z · score: 1 (6 votes)
Are causal decision theorists trying to outsmart conditional probabilities? 2017-05-16T08:01:27.426Z · score: 4 (5 votes)
Publication on formalizing preference utilitarianism in physical world models 2015-09-22T16:46:54.934Z · score: 5 (6 votes)
Two-boxing, smoking and chewing gum in Medical Newcomb problems 2015-06-29T10:35:58.162Z · score: 15 (18 votes)
Request for feedback on a paper about (machine) ethics 2014-09-28T12:03:05.500Z · score: 7 (8 votes)

Comments

Comment by caspar42 on Pavlov Generalizes · 2019-05-17T00:16:04.817Z · score: 5 (3 votes) · LW · GW

Not super important but maybe worth mentioning in the context of generalizing Pavlov: the strategy Pavlov for the iterated PD can be seen as an extremely shortsighted version of the law of effect, which basically says: repeat actions that have worked well in the past (in similar situations). Of course, the LoE can be applied in a wide range of settings. For example, in their reinforcement learning textbook, Sutton and Barto write that LoE underlies all of (model-free) RL.

Comment by caspar42 on In memoryless Cartesian environments, every UDT policy is a CDT+SIA policy · 2019-01-16T23:54:51.593Z · score: 1 (1 votes) · LW · GW

Elsewhere, I illustrate this result for the absent-minded driver.

Comment by caspar42 on CDT=EDT=UDT · 2019-01-16T23:50:26.486Z · score: 11 (3 votes) · LW · GW

> I tried to understand Caspar’s EDT+SSA but was unable to figure it out. Can someone show how to apply it to an example like the AMD to help illustrate it?

Sorry about that! I'll try to explain it some more. Let's take the original AMD. Here, the agent only faces a single type of choice -- whether to EXIT or CONTINUE. Hence, in place of a policy we can just condition on when computing our SSA probabilities. Now, when using EDT+SSA, we assign probabilities to being a specific instance in a specific possible history of the world. For example, we assign probabilities of the form , which denotes the probability that given I choose to CONTINUE with probability , history (a.k.a. CONTINUE, EXIT) is actual and that I am the instance intersection (i.e., the first intersection). Since we're using SSA, these probabilities are computed as follows:

That is, we first compute the probability that the history itself is actual (given ). Then we multiply it by the probability that within that history I am the instance at , which is just 1 divided by the number of instances of myself in that history, i.e. 2.

Now, the expected value according to EDT + SSA given can be computed by just summing over all possible situations, i.e. over all combinations of a history and a position within that history and multiplying the probability of that situation with the utility given that situation:

And that's exactly the ex ante expected value (or UDT-expected value, I suppose) of continuing with probability . Hence, EDT+SSA's recommendation in AMD is the ex ante optimal policy (or UDT's recommendation, I suppose). This realization is not original to myself (though I came up with it independently in collaboration with Johannes Treutlein) -- the following papers make the same point:

  • Rachael Briggs (2010): Putting a value on Beauty. In Tamar Szabo Gendler and John Hawthorne, editors, Oxford Studies in Epistemology: Volume 3, pages 3–34. Oxford University Press, 2010. http://joelvelasco.net/teaching/3865/briggs10-puttingavalueonbeauty.pdf
  • Wolfgang Schwarz (2015): Lost memories and useless coins: revisiting the absentminded driver. In: Synthese. https://www.umsu.de/papers/driver-2011.pdf

My comment generalizes these results a bit to include cases in which the agent faces multiple different decisions.

Comment by caspar42 on CDT Dutch Book · 2019-01-16T22:37:51.098Z · score: 11 (3 votes) · LW · GW
Caspar Oesterheld is working on similar ideas.

For anyone who's interested, Abram here refers to my work with Vincent Conitzer which we write about here.

Comment by caspar42 on Reflexive Oracles and superrationality: prisoner's dilemma · 2018-11-26T00:28:57.072Z · score: 1 (1 votes) · LW · GW

My paper "Robust program equilibrium" (published in Theory and Decision) discusses essentially NicerBot (under the name ϵGroundedFairBot) and mentions Jessica's comment in footnote 3. More generally, the paper takes strategies from iterated games and transfers them into programs for the corresponding program game. As one example, tit for tat in the iterated prisoner's dilemma gives rise to NicerBot in the "open-source prisoner's dilemma".

Comment by caspar42 on Naturalized induction – a challenge for evidential and causal decision theory · 2018-05-29T14:48:23.686Z · score: 3 (1 votes) · LW · GW

Link to relevant agent foundations forum comment

Comment by caspar42 on Idea: OpenAI Gym environments where the AI is a part of the environment · 2018-04-14T18:27:26.760Z · score: 3 (1 votes) · LW · GW

I list some relevant discussions of the "anvil problem" etc. here. In particular, Soares and Fallenstein (2014) seem to have implemented an environment in which such problems can be modeled.

Comment by caspar42 on Announcement: AI alignment prize winners and next round · 2018-03-28T12:19:44.214Z · score: 3 (1 votes) · LW · GW

For this round I submit the following entries on decision theory:

Robust Program Equilibrium (paper)

The law of effect, randomization and Newcomb’s problem (blog post) (I think James Bell's comment on this post makes an important point.)

A proof that every ex-ante-optimal policy is an EDT+SSA policy in memoryless POMPDs (IAFF comment) (though see my own comment to that comment for a caveat to that result)

Comment by caspar42 on Causal Universes · 2018-02-16T17:09:26.439Z · score: 0 (0 votes) · LW · GW

(RobbBB seems to refer to what philosophers call the B-theory of time, whereas CronoDAS seems to refer to the A-theory of time.)

Comment by caspar42 on In memoryless Cartesian environments, every UDT policy is a CDT+SIA policy · 2018-02-11T19:10:10.000Z · score: 13 (5 votes) · LW · GW

Since Briggs [1] shows that EDT+SSA and CDT+SIA are both ex-ante-optimal policies in some class of cases, one might wonder whether the result of this post transfers to EDT+SSA. I.e., in memoryless POMDPs, is every (ex ante) optimal policy also consistent with EDT+SSA in a similar sense. I think it is, as I will try to show below.

Given some existing policy , EDT+SSA recommends that upon receiving observation we should choose an action from (For notational simplicity, I'll assume that policies are deterministic, but, of course, actions may encode probability distributions.) Here, if and otherwise. is the SSA probability of being in state of the environment trajectory given the observation and the fact that one uses the policy .

The SSA probability is zero if and otherwise. Here, is the number of times occurs in . Note that this is the minimal reference class version of SSA, also known as the double-halfer rule (because it assigns 1/2 probability to tails in the Sleeping Beauty problem and sticks with 1/2 if it's told that it's Monday).

Inserting this into the above, we get where the first sum on the right-hand side is over all histories that give rise to observation at some point. Dividing by the number of agents with observation in a history and setting the policy for all agents at the same time cancel each other out, such that this equals Obviously, any optimal policy chooses in agreement with this. But the same disclaimers apply; multiple policies satisfy the right-hand side of this equation and not all of these are optimal.

[1] Rachael Briggs (2010): Putting a value on Beauty. In Tamar Szabo Gendler and John Hawthorne, editors, Oxford Studies in Epistemology: Volume 3, pages 3–34. Oxford University Press, 2010. http://joelvelasco.net/teaching/3865/briggs10-puttingavalueonbeauty.pdf

Comment by caspar42 on In memoryless Cartesian environments, every UDT policy is a CDT+SIA policy · 2018-02-11T19:09:11.000Z · score: 10 (3 votes) · LW · GW

Caveat: The version of EDT provided above only takes dependences between instances of EDT making the same observation into account. Other dependences are possible because different decision situations may be completely "isomorphic"/symmetric even if the observations are different. It turns out that the result is not valid once one takes such dependences into account, as shown by Conitzer [2]. I propose a possible solution in https://casparoesterheld.com/2017/10/22/a-behaviorist-approach-to-building-phenomenological-bridges/ . Roughly speaking, my solution is to identify with all objects in the world that are perfectly correlated with you. However, the underlying motivation is unrelated to Conitzer's example.

[2] Vincent Conitzer: A Dutch Book against Sleeping Beauties Who Are Evidential Decision Theorists. Synthese, Volume 192, Issue 9, pp. 2887-2899, October 2015. https://arxiv.org/pdf/1705.03560.pdf

Comment by caspar42 on Prisoner's dilemma tournament results · 2018-02-02T16:03:03.186Z · score: 1 (1 votes) · LW · GW

I tried to run this with racket and #lang scheme (as well as #lang racket) but didn't get it to work (though I didn't try for very long), perhaps because of backward compatibility issues. This is a bit unfortunate because it makes it harder for people interested in this topic to profit from the results and submitted programs of this tournament. Maybe you or Alex could write a brief description of how one could get the program tournament to run?

Comment by caspar42 on The Absent-Minded Driver · 2018-01-26T11:13:44.499Z · score: 3 (1 votes) · LW · GW

I wonder what people here think about the resolution proposed by Schwarz (2014). His analysis is that the divergence from the optimal policy also goes away if one combines EDT with the halfer position a.k.a. the self-sampling assumption, which, as shown by Briggs (2010), appears to be the right anthropic view to combine with EDT, anyway.

Comment by caspar42 on A model I use when making plans to reduce AI x-risk · 2018-01-20T10:14:15.591Z · score: 21 (9 votes) · LW · GW

I think this is a good overview, but most of the views proposed here seem contentious and the arguments given in support shouldn't suffice to change the mind of anyone who has thought about these questions for a bit or who is aware of the disagreements about them within the community.

Getting alignment right accounts for most of the variance in whether an AGI system will be positive for humanity.

If your values differ from those of the average human, then this may not be true/relevant. E.g., I would guess that for a utilitarian current average human values are worse than, e.g., 90% "paperclipping values" and 10% classical utilitarianism.

Also, if gains from trade between value systems are big, then a lot of value may come from ensuring that the AI engages in acausal trade (https://wiki.lesswrong.com/wiki/Acausal_trade ). This is doubly persuasive if you already see your own policies as determining what agents with similar decision theories but different values do elsewhere in the universe. (See, e.g., section 4.6.3 of "Multiverse-wide Cooperation via Correlated Decision Making".)

Given timeline uncertainty, it's best to spend marginal effort on plans that assume / work in shorter timelines.
Stated simply: If you don't know when AGI is coming, you should make sure alignment gets solved in worlds where AGI comes soon.

I guess the question is what "soon" means. I agree with the argument provided in the quote. But there are also some arguments to work on longer timelines, e.g.:

  • If it's hard and most value comes from full alignment, then why even try to optimize for very short timelines?
  • Similarly, there is a "social" difficulty of getting people in AI to notice your (or the AI safety community's) work. Even if you think you could write down within a month a recipe for increasing the probability of AI being aligned by a significant amount, you would probably need much more than a month to make it significantly more likely to get people to consider applying your recipe.

It seems obvious that most people shouldn't think too much about extremely short timelines (<2 years) or the longest plausible timelines (>300 years). So, these arguments together probably point to something in the middle of these and the question is where. Of course, it also depends on one's beliefs about AI timelines.

To me it seems that the concrete recommendations (aside from the "do AI safety things") don't have anything to do with the background assumptions.

As one datapoint, fields like computer science, engineering and mathematics seem to make a lot more progress than ones like macroeconomics, political theory, and international relations.

For one, "citation needed". But also: the alternative to doing technical AI safety work isn't to do research in politics but to do political activism (or lobbying or whatever), i.e. to influence government policy.

As your "technical rather than political" point currently stands, it's applicable to any problem, but it is obviously invalid at this level of generality. To argue plausibly that technical work on AI safety is more important than AI strategy (which is plausibly true), you'd have to refer to some specifics of the problems related to AI.

Comment by caspar42 on Prediction Markets are Confounded - Implications for the feasibility of Futarchy · 2018-01-16T14:20:22.938Z · score: 0 (0 votes) · LW · GW

The issue with this example (and many similar ones) is that to decide between interventions on a variable X from the outside, EDT needs an additional node representing that outside intervention, whereas Pearl-CDT can simply do(X) without the need for an additional variable. If you do add these variables, then conditioning on that variable is the same as intervening on the thing that the variable intervenes on. (Cf. section 3.2.2 "Interventions as variables" in Pearl's Causality.)

Comment by caspar42 on Niceness Stealth-Bombing · 2018-01-14T11:48:39.364Z · score: 3 (1 votes) · LW · GW

This advice is very similar to Part, 1, ch. 3; Part 3, ch. 5; Part 4, ch. 1, 6 in Dale Carnegie's classic How to Win Friends and Influence People.

Comment by caspar42 on Writing Down Conversations · 2017-12-31T08:50:52.069Z · score: 11 (3 votes) · LW · GW

Another classic on this topic by a community member is Brian Tomasik's Turn Discussions Into Blog Posts.

Comment by caspar42 on The expected value of the long-term future · 2017-12-30T23:40:40.615Z · score: 22 (6 votes) · LW · GW

I looked at the version 2017-12-30 10:48:11Z.

Overall, I think it's a nice, systematic overview. Below are some comments.

I should note that I'm not very expert on these things. This is also why the additional literature I mention is mostly weakly related stuff from FRI, the organization I work for. Sorry about that.

An abstract would be nice.

Locators in the citations would be useful, i.e. "Beckstead (2013, sect. XYZ)" instead of just "Beckstead (2013)" when you talk about some specific section of the Beckstead paper. (Cf. section “Pageless Documentation” of the humurous Academic Citation Practice: A Sinking Sheep? by Ole Bjørn Rekdal.)

>from a totalist, consequentialist, and welfarist (but not necessarily utilitarian) point of view

I don't think much of your analysis assumes welfarism (as I understand it)? Q_w could easily denote things other than welfare (e.g., how virtue ethical, free, productive, autonomous, natural, the mean person is), right? (I guess some of the discussion sections are fairly welfarist, i.e. they talk about suffering, etc., rather than freedom and so forth.)

>an existential risk as one where an adverse outcome would either annihilate Earth-originating intelligent life or permanently and drastically curtail its potential.

Maybe some people would interpret this definition as excluding some of the "shrieks" and "whimpers", since in some of them, "humanity's potential is realized" in that it colonizes space, but not in accordance with, e.g., the reader's values. Anyway, I think this definition is essentially a quote from Bostrom (maybe use quotation marks?), so it's alright.

>The first is the probability P of reaching time t.

Maybe say more about why you separate N_w(t) (in the continuous model) into P(t) and N(t)?

I also don't quite understand whether equation 1 is intended as the expected value of the future or as the expected value of a set of futures w that all have the same N_w(t) and Q_w(t). The problem is that if it's the expected value of the future, I don't get how you can simplify something like

into the right side of your equation 1. (E.g., you can't just let N(t) and Q(t) denote expected numbers of moral patients and expected mean qualities of life, because the mean qualities in larger worlds ought to count for more, right?)

I suspect that when reading the start of sect. 3.1, a lot of readers will wonder whether you endorse all the assumptions underlying your model of P(t). In particular, I would guess that people would disagree with the following two assumptions:

-> Short term x-risk reduction (r_1) doesn't have any effect on long-term risk (r). Perhaps this is true for some fairly specific work on preventing extinction but it seems less likely for interventions like building up the UN (to avoid all kinds of conflict, coordinate against risks, etc.).

-> Long-term extinction risk is constant. I haven't thought much about these issues but I would guess that extinction risk becomes much lower, once there is a self-sustaining colony on Mars.

Reading further, I see that you address these in sections 3.2 and 3.3. Maybe you could mention/refer to these somewhere near the start of sect. 3.1.

On page 3, you say that the derivative of -P(t) w.r.t. r_1 denotes the value of reducing r_1 by one unit. This is true in this case because P(t) is linear in r_1. But in general, the value of reducing r_1 by one unit is just P(t,r_1-1)-P(t,r_1), right?

Is equation 3, combined with the view that the cost of one unit of f1 is constant, consistent with Ord's "A plausible model would be that it is roughly as difficult to halve the risk per century, regardless of its starting probability, and more generally, that it is equally difficult to reduce it by some proportion regardless of its absolute value beforehand."? With your model, it looks like bringing f_1 from 0 to 0.5 and thus halfing r_1 is just as expensive as bringing f_1 from 0.5 to 1.

On p. 7, "not to far off" -- probably you mean "too"?

>For example, perhaps we will inevitably develop some hypothetical weapons that give so large an advantage to offence over defence that civilisation is certain to be destroyed.

AI risk is another black ball that will become more accessible. But maybe you would rather not model it as extinction. At least AI risk doesn't necessarily explain the Fermi paradox and AIs may create sentient beings.

>Ord argues that we may be able to expect future generations to be more interested in risk reduction, implying increasing f_i

I thought f_i was meant to model the impact that we can have on r_i? So, to me it seems more sensible to model the involvement of future generations, to the extent that we can't influence it, as a "a kind of event E" (as you propose) or, more generally, as implying that the non-intervention risk levels r_i decrease.

>This would only reinforce the case for extinction risk reduction.

It seems that future generations caring about ERR makes short-term ERR more important (because the long-term future is longer and thus can contain more value). But it makes long-term ERR less important, because future generations will, e.g., do AI safety research anyway. (In section "Future resources" of my blog post Complications in evaluating neglectedness, I make the general point that for evaluating the neglectedness of an intervention, one has to look at how many resources future generations will invest into that intervention.)

>There is one case in which it clearly is not: if space colonisation is in fact likely to involve risk-independent islands. Then high population goes with low risk, increasing the value of the future relative to the basic model

(I find risk-independent islands fairly plausible.)

>The expected number of people who will live in period t is

You introduced N(t) as the number of morally relevant beings (rather than "people").

>However, this increase in population may be due to stop soon,

Although it is well-known that some predict population to stagnate at 9 billion or so, a high-quality citation would be nice.

>The likelihood of space colonisation, a high-profile issue on which billions of dollars is spent per year (Masters, 2015), also seems relatively hard to affect. Extinction risk reduction, on the other hand, is relatively neglected (Bostrom, 2013; Todd, 2017), so it could be easier to achieve progress in this area.

I have only briefly (in part due to the lack of locators) checked the two sources, but it seems that this varies strongly between different extinction risks. For instance, according to Todd (2017), >300bn (and thus much more than on space colonization) is spent on climate change, 1-10bn on nuclear security, 1bn on extreme pandemic prevention. So, overall much more money goes into extinction risk reduction than into space colonization. (This is not too surprising. People don't want to die, and they don't want their children or grandchildren to die. They don't care nearly as much about whether some elite group of people will live on Mars in 50 years.)

Of course, there a lot of complications to this neglectedness analysis. (All three points I discuss in Complications in evaluating neglectedness seem to apply.)

>Some people believe that it’s nearly impossible to have a consistent impact on Q(t) so far into the future.

Probably a reference would be good. I guess to the extent that we can't affect far future Q(t), we also can't affect far future r_i.

>However, this individual may be biased against ending things, for instance because of the survival instinct, and so could individuals or groups in the future. The extent of this bias is an open question.

It's also a bit unclear (at least based on hat you write) what legitimizes calling this a bias, rather than simply a revealed preference not to die (even in cases in which you or I as outside observers might think it to be preferable not to live) and thus evidence that their lives are positive. Probably one has to argue via status quo bias or sth like that.

>We may further speculate that if the future is controlled by altruistic values, even powerless persons are likely to have lives worth living. If society is highly knowledgeable and technologically sophisticated, and decisions are made altruistically, it’s plausible that many sources of suffering would eventually be removed, and no new ones created unnecessarily. Selfish values, on the other hand, do not care about the suffering of powerless sentients.

This makes things sound a more binary than they actually are. (I'm sure you're aware of this.) In the usual sense of the word, people could be "altruistic" but in a non-consequentialist way. There may be lots of suffering in such worlds. (E.g., some libertarians may be regard intervening in the economy as unethical even if companies start creating slaves. A socialist, on the other hand, may view capitalism as fundamentally unjust, try to regulate/control the economy and thus cause a lot of poverty.) Also, even if someone is altruistic in a fairly consequentialist way, they may still not care about all beings that you/I/the reader cares about. E.g., economists tend to be consequentialists but rarely consider animal welfare.

I think for the animal suffering (both wild animals and factory farming) it is worth noting that it seems fairly unlikely that this will be economically efficient in the long term, but that the general underlying principles (Darwinian suffering and exploiting the powerless) might carry over to other beings (like sentient AIs).

Another way in which the future may be negative would be the Malthusian trap btw. (Of course, some would regard at least some Malthusian trap scenarios as positive, see, e.g., Robin Hanson's The Age of Em.) Presumably this belongs to 5.2.1, since it's a kind of coordination failure.

As you say, I think the option value argument isn't super persuasive, because it seems unlikely that the people in power in a million years share my (meta-)values (or agree with the way I do compromise).

Re 5.2.3: Another relevant reference on why one should cooperate -- which is somewhat separate from the point that if mutual cooperation works out the gains from trade are great -- is Brian Tomasik's Reasons to Be Nice to Other Value Systems.

>One way to increase Q(t) is to advocate for positive value changes in the direction of greater consideration for powerless sentients, or to promote moral enhancement (Persson and Savulescu, 2008). Another approach might be to work to improve political stability and coordination, making conflict less likely as well as increasing the chance that moral progress continues.

Relevant:

https://foundational-research.org/international-cooperation-vs-ai-arms-race/

http://reducing-suffering.org/values-spreading-often-important-extinction-risk/

Comment by caspar42 on Announcing the AI Alignment Prize · 2017-12-21T16:10:18.655Z · score: 15 (4 votes) · LW · GW

You don't mention decision theory in your list of topics, but I guess it doesn't hurt to try.

I have thought a bit about what one might call the "implementation problem of decision theory". Let's say you believe that some theory of rational decision making, e.g., evidential or updateless decision theory, is the right one for an AI to use. How would you design an AI to behave in accordance with such a normative theory? Conversely, if you just go ahead and build a system in some existing framework, how would that AI behave in Newcomb-like problems?

There are two pieces that I uploaded/finished on this topic in November and December. The first is a blog post noting that futarchy-type architectures would, per default, implement evidential decision theory. The second is a draft titled "Approval-directed agency and the decision theory of Newcomb-like problems".

For anyone who's interested in this topic, here are some other related papers and blog posts:

So far, my research and the papers by others I linked have focused on classic Newcomb-like problems. One could also discuss how existing AI paradigms related to other issues of naturalized agency, in particular self-locating beliefs and naturalized induction, though here it seems more as though existing frameworks just lead to really messy behavior.

Send comments to firstnameDOTlastnameATfoundational-researchDOTorg. (Of course, you can also comment here or send you a LW PM.)

Comment by caspar42 on Superintelligence via whole brain emulation · 2017-12-17T08:49:34.764Z · score: 0 (0 votes) · LW · GW

I wrote a summary of Hansons's The Age of Em, in which I focus on the bits of information that may be policy-relevant for effective altruists. For instance, I summarize what Hanson says about em values and also have a section about AI safety.

Comment by caspar42 on Intellectual Hipsters and Meta-Contrarianism · 2017-11-11T08:56:34.630Z · score: 0 (0 votes) · LW · GW

Great post, obviously.

You argue that signaling often leads to distribution of intellectual positions following this pattern: in favor of X with simple arguments / in favor of Y with complex arguments / in favor of something like X with simple arguments

I think it’s worth noting that the pattern of position often looks different. For example, there is: in favor of X with simple arguments / in favor of Y with complex arguments / in favor of something like X with surprising and even more sophisticated and hard-to-understand arguments

In fact, I think many of your examples follow the latter pattern. For example, the market efficiency arguments in favor of libertarianism seem harder-to-understand and more sophisticated than most arguments for liberalism. Maybe it fits your pattern better if libertarianism is justified purely on the basis of expert opinion.

Similarly, the justification for the “meta-contrarian” position in "don't care about Africa / give aid to Africa / don't give aid to Africa" is more sophisticated than the reasons for the contrarian or naive positions.

But as has been pointed out, along with the gigantic cost, death does have a few small benefits. It lowers overpopulation, it allows the new generation to develop free from interference by their elders, it provides motivation to get things done quickly.

I’m not sure whether the overpopulation is a good example. I think in many circles that point would signal naivety and people would respond by something deep-sounding about how life is sacred. (The same is true for “it’s good if old people die because that saves money and allows the government to build more schools”.) Here, too, I would argue that your pattern doesn’t quite describe the set of commonly held positions, as it omits the naive pro-death position.

Comment by caspar42 on Are causal decision theorists trying to outsmart conditional probabilities? · 2017-10-06T16:00:55.732Z · score: 0 (0 votes) · LW · GW

I agree that in situations where A only has outgoing arrows, p(s | do(a)) = p(s | a), but this class of situations is not the "Newcomb-like" situations.

What I meant to say is that the situations where A only has outgoing arrows are all not Newcomb-like.

Maybe we just disagree on what "Newcomb-like" means? To me what makes a situation "Newcomb-like" is your decision algorithm influencing the world through something other than your decision (as happens in the Newcomb problem via Omega's prediction). In smoking lesion, this does not happen, your decision algorithm only influences the world via your action, so it's not "Newcomb-like" to me.

Ah, okay. Yes, in that case, it seems to be only a terminological dispute. As I say in the post, I would define Newcomb-like-ness via a disagreement between EDT and CDT which can mean either that they disagree about what the right decision is, or, more naturally, that their probabilities diverge. (In the latter case, the statement you commented on is true by definition and in the former case it is false for the reason I mentioned in my first reply.) So, I would view the Smoking lesion as a Newcomb-like problem (ignoring the tickle defense).

Comment by caspar42 on Principia Compat. The potential Importance of Multiverse Theory · 2017-10-06T14:59:01.055Z · score: 0 (0 votes) · LW · GW

Yes, the paper is relatively recent, but in May I published a talk on the same topic. I also asked on LW whether someone would be interested in giving feedback a month or so before actually the paper.

Do you think your proof/argument is also relevant for my multiverse-wide superrationality proposal?

Comment by caspar42 on Are causal decision theorists trying to outsmart conditional probabilities? · 2017-10-06T14:53:44.219Z · score: 0 (0 votes) · LW · GW

So, the class of situations in which p(s | do(a)) = p(s | a) that I was alluding to is the one in which A has only outgoing arrows (or all the values of A’s predecessors are known). (I guess this could be generalized to: p(s | do(a)) = p(s | a) if A d-separates its predecessors from S?) (Presumably this stuff follows from Rule 2 of Theorem 3.4.1 in Causality.)

All problems in which you intervene in an isolated system from the outside are of this kind and so EDT and CDT make the same recommendations for intervening in a system from the outside. (That’s similar to the point that Pearl makes in section 3.2.2 of Causality: You can model the do-interventions by adding action nodes without predecessors and conditioning on these action nodes.)

The Smoking lesion is an example of a Newcomb-like problem where A has an inbound arrow that leads p(s | do(a)) and p(s | a) to differ. (That said, I think the smoking lesion does not actually work as a Newcomb-like problem, see e.g. chapter 4 of Arif Ahmed’s Evidence, Decision and Causality.)

Similarly, you could model Newcomb’s problem by introducing a logical node as a predecessor of your decision and the result of the prediction. (If you locate “yourself” in the logical node and the logical node does not have any predecessors, then CDT and EDT agree again.)

Of course, in the real world, all problems are in theory Newcomb-like because there are always some ingoing arrows into your decision. But in practice, most problems are nearly non-Newomb-like because, although there may be an unblocked path from my action to the value of my utility function, that path is usually too long/complicated to be useful. E.g., if I raise my hand now, that would mean that the state of the world 1 year ago was such that I raise my hand now. And the world state 1 year ago causes how much utility I have. But unless I’m in Arif Ahmed’s “Betting on the Past”, I don’t know which class of world states 1 year ago (the ones that lead to me raising my hand or the ones that cause me not to raise my hand) causes me to have more utility. So, EDT couldn't try to exploit that way of changing the past.

Comment by caspar42 on A survey of polls on Newcomb’s problem · 2017-09-28T07:15:02.673Z · score: 0 (0 votes) · LW · GW

I claim that one-boxers do not believe b and c are possible because Omega is cheating or a perfect predictor (same thing)

Note that Omega isn't necessarily a perfect predictor. Most one-boxers would also one-box if Omega is a near-perfect predictor.

Aside from "lizard man", what are the other reasons that lead to two-boxing?

I think I could pass an intellectual Turing test (the main arguments in either direction aren't very sophisticated), but maybe it's easiest to just read, e.g., p. 151ff. of James Joyce's The Foundations of Causal Decision Theory and note how Joyce understands the problem in pretty much the same way that a one-boxer would.

In particular, Joyce agrees that causal decision theorists would want to self-modify to become one-boxers. (I have heard many two-boxers admit to this.) This doesn't make sense if they don't believe in Omega's prediction abilities.

Comment by caspar42 on Naturalized induction – a challenge for evidential and causal decision theory · 2017-09-28T06:58:23.514Z · score: 0 (0 votes) · LW · GW

I hadn’t seen these particular discussions, although I was aware of the fact that UDT and other logical decision theories avoid building phenomenological bridges in this way. I also knew that others (e.g., the MIRI people) were aware of this.

I didn't know you preferred a purely evidential variant of UDT. Thanks for the clarification!

As for the differences between LZEDT and UDT:

  • My understanding was that there is no full formal specification of UDT. The counterfactuals seem to be given by some unspecified mathematical intuition module. LZEDT, on the other hand, seems easy to specify formally (assuming a solution to naturalized induction). (That said, if UDT is just the updateless-evidentialist flavor of logical decision theory, it should be easy to specify as well. I haven’t seen people UDT characterize in this way, but perhaps this is because MIRI’s conception of UDT differs from yours?)
  • LZEDT isn’t logically updateless.
  • LZEDT doesn’t do explicit optimization of policies. (Explicit policy optimization is the difference between UDT1.1 and UDT1.0, right?)

(Based on a comment you made on an earlier past post of mine, it seems that UDT and LZEDT reason similarly about medical Newcomb problems.)

Anyway, my reason for writing this isn’t so much that LZEDT differs from other decision theories. (As I say in the post, I actually think LZEDT is equivalent to the most natural evidentialist logical decision theory — which has been considered by MIRI at least.) Instead, it’s that I have a different motivation for proposing it. My understanding is that the LWers’ search for new decision theories was not driven by the BPB issue (although some of the motivations you listed in 2012 are related to it). Instead it seems that people abandoned EDT — the most obvious approach — mainly for reasons that I don’t endorse. E.g., the TDT paper seems to give medical Newcomb problems as the main argument against EDT. It may well be that looking beyond EDT to avoid naturalized induction/BPB leads to the same decision theories as these other motivations.

Comment by caspar42 on Naturalized induction – a challenge for evidential and causal decision theory · 2017-09-25T23:55:56.610Z · score: 1 (1 votes) · LW · GW

Yes, I share the impression that the BPB problem implies some amount of decision theory relativism. That said, one could argue that decision theories cannot be objectively correct, anyway. In most areas, statements can only be justified relative to some foundation. Probability assignments are correct relative to a prior, the truth of theorems depends on axioms, and whether you should take some action depends on your goals (or meta-goals). Priors, axioms, and goals themselves, on the other hand, cannot be justified (unless you have some meta-priors, meta-axioms, etc., but I think the chain as to end at some point, see https://en.wikipedia.org/wiki/Regress_argument ). Perhaps decision theories are similar to priors, axioms and terminal values?

Comment by caspar42 on Naturalized induction – a challenge for evidential and causal decision theory · 2017-09-23T06:06:41.620Z · score: 1 (1 votes) · LW · GW

No, I actually mean that world 2 doesn't exist. In this experiment, the agent believes that either world 1 or world 2 is actual and that they cannot be actual at the same time. So, if the agent thinks that it is in world 1, world 2 doesn't exist.

Comment by caspar42 on Are causal decision theorists trying to outsmart conditional probabilities? · 2017-09-23T04:01:51.370Z · score: 0 (0 votes) · LW · GW

(Sorry again for being slow to reply to this one.)

"Note that in non-Newcomb-like situations, P(s|do(a)) and P(s|a) yield the same result, see ch. 3.2.2 of Pearl’s Causality."

This is trivially not true.

Is this because I define "Newcomb-ness" via disagreement about the best action between EDT and CDT in the second paragraph? Of course, the d(P(s|do(a)),P(s|a)) could be so small that EDT and CDT agree on what action to take. They could even differ in such a way that CDT-EV and EDT-EV are the same.

But it seems that instead of comparing the argmaxes or the EVs, one could also use the term Newcomb-ness on the basis of the probabilities themselves. Or is there some deeper reason why the sentence is false?

Comment by caspar42 on Naturalized induction – a challenge for evidential and causal decision theory · 2017-09-23T03:52:29.501Z · score: 3 (3 votes) · LW · GW

I apologize for not replying to your earlier comment. I do engage with comments a lot. E.g., I recall that your comment on that post contained a link to a ~1h talk that I watched after reading it. There are many obvious reasons that sometimes cause me not reply to comments, e.g. if I don't feel like I have anything interesting to say, or if the comment indicates lack of interest in discussion (e.g., your "I am not actually here, but ... Ok, disappearing again"). Anyway, I will reply your comment now. Sorry again for not doing so earlier.

Comment by caspar42 on Naturalized induction – a challenge for evidential and causal decision theory · 2017-09-22T20:57:59.260Z · score: 1 (1 votes) · LW · GW

I just remembered that in Naive TDT, Bayes nets, and counterfactual mugging, Stuart Armstrong made the point that it shouldn't matter whether you are simulated (in a way that you might be the simulation) or just predicted (in such a way that you don't believe that you could be the simulation).

Comment by caspar42 on Splitting Decision Theories · 2017-09-22T20:45:27.038Z · score: 2 (2 votes) · LW · GW

Interesting post! :)

I think the process is hard to formalize because specifying step 2 seems to require specifying a decision theory almost directly. Recall that causal decision theorists argue that two-boxing is the right choice in Newcomb’s problem. Similarly, some would argue that not giving the money in counterfactual mugging is the right choice from the perspective of the agent who already knows that it lost, whereas others argue for the opposite. Or take a look at the comments on the Two-Boxing Gene. Generally, the kind of decision problems that put decision theories to a serious test also tend to be ones in which it is non-obvious what the right choice is. The same applies to meta-principles. Perhaps people agree with the vNM axioms, but desiderata that could shed a light on Newcomblike problems appear to be more controversial. For example, irrelevance of impossible outcomes and reflective stability both seem desirable but actually contradict each other.

TL;DR: It seems to be really hard to specify what it means for a decision procedure to "win"/fail in a given thought experiment.

Comment by caspar42 on A survey of polls on Newcomb’s problem · 2017-09-21T20:37:22.915Z · score: 0 (0 votes) · LW · GW

Yeah, I also think the "fooling Omega idea" is a common response. Note however that two-boxing is more common among academic decision theorists, all of which understand that Newcomb's problem is set up such that you can't fool Omega. I also doubt that the fooling Omega idea is the only (or even the main) cause of two-boxing among non-decision theorists.

Comment by caspar42 on Multiverse-wide Cooperation via Correlated Decision Making · 2017-08-31T16:15:18.815Z · score: 0 (0 votes) · LW · GW

Thanks for the comment!

W.r.t. moral reflection: Probably many agents put little intrinsic value on whether society engages in a lot of moral reflection. However, I would guess that as a whole the set of agents having a similar decision mechanism as I have do care about this significantly and positively. (Empirically, disvaluing moral reflection seems to be rare.) Hence, (if the basic argument of the paper goes through) I should give some weight to it.

W.r.t. moral pluralism: Probably even fewer agents care about this intrinsically. I certainly don’t care about it intrinsically. The idea is that moral pluralism may avoid conflict or create gains from “trade”. For example, let’s say the aggregated values of agents with my decision algorithm contain two values A and B. (As I argue in the paper, I should maximize these aggregated values to maximize my own values throughout the multiverse.) Now, I might be in some particular environment with agents who themselves care about A and/or B. Let’s say I can choose between two distributions of caring about A and B: Either each of the agents cares about A and B, or some care only about A and the others only about B. The former will tend to be better if I (or rather the set of agents with my decision algorithm) care about A and B, because it avoids conflicts, makes it more easy to exploit comparative advantages, etc.

Note that I think neither promoting moral reflection nor promoting moral pluralism is a strong candidate for a top intervention. Multiverse-wide superrationality just increases their value relative to what, say, what a utilitarian would think about these interventions. I think it’s a lot more important to ensure that AI uses the right decision theory. (Of course, this is important, anyway, but I think multiverse-wide superrationality drastically increases its value.)

Comment by caspar42 on Principia Compat. The potential Importance of Multiverse Theory · 2017-08-26T08:12:30.437Z · score: 2 (2 votes) · LW · GW

I recently published a different proposal for implementing acausal trade as humans: https://foundational-research.org/multiverse-wide-cooperation-via-correlated-decision-making/ Basically, if you care about other parts of the universe/multiverse and these parts contain agents that are decision-theoretically similar to you, you can cooperate with them via superrationality. For example, let's say I give most moral weight to utilitarian considerations and care less about, e.g., justice. Probably other parts of the universe contain agents that reason about decision theory in the same way that I do. Because of orthogonality ( https://wiki.lesswrong.com/wiki/Orthogonality_thesis ), many of these will have other goals, though most of them will probably have goals that arise from evolution. Then if I expect (based on the empirical study of humans or thinking about evolution) that many other agents care a lot about justice, this gives me a reason to give more weight to justice as this makes it more likely (via superrationality / EDT / TDT / ... ) that other agents also give more weight to my values.

Comment by caspar42 on Invitation to comment on a draft on multiverse-wide cooperation via alternatives to causal decision theory (FDT/UDT/EDT/...) · 2017-06-23T16:42:12.958Z · score: 1 (1 votes) · LW · GW

Yep, this is roughly the type of cooperation I have in mind. Some minor points:

Overall, I am not sure whether gains from trade would arise in this specific scenario. Perhaps, it’s not better for the civilizations than if each civilization only builds habitats for itself?

The game theoretic motive is that, by doing something good for a hypothetical species, there might exist an inaccessible universe in which that species is both living and able to surmise that the humans have done this, and that they will by luck create a small utopia of humans when they do their counterpart project.

As I argue in section “No reciprocity needed: whom to treat beneficially”, the benefit doesn’t necessarily come from the species that we benefit. Even if agent X is certain that agent Y cannot benefit X, agent X may still help Y to make it more likely that X receives help from other agents who are in a structurally similar situation w.r.t. Y and think about it in a way similar to X.

Also, the other civilizations don’t need to be able to check whether we helped them, just like in the prisoner’s dilemma against a copy we don’t have to be able to check whether the other copy actually cooperated. It’s enough to know, prior to making one’s own decision, that the copy reasons similarly about these types of problems.

Comment by caspar42 on Invitation to comment on a draft on multiverse-wide cooperation via alternatives to causal decision theory (FDT/UDT/EDT/...) · 2017-06-23T16:39:59.844Z · score: 0 (0 votes) · LW · GW

Yep, this is roughly the type of cooperation I have in mind. Some minor points:

Overall, I am not sure whether gains from trade would arise in this specific scenario. Perhaps, it’s not better for the civilizations than if each civilization only builds habitats for itself?

The game theoretic motive is that, by doing something good for a hypothetical species, there might exist an inaccessible universe in which that species is both living and able to surmise that the humans have done this, and that they will by luck create a small utopia of humans when they do their counterpart project.

As I argue in section “No reciprocity needed: whom to treat beneficially”, the benefit doesn’t necessarily come from the species that we benefit. Even if agent X is certain that agent Y cannot benefit X, agent X may still help Y to make it more likely that X receives help from other agents who are in a structurally similar situation w.r.t. Y and think about it in a way similar to X.

Also, the other civilizations don’t need to be able to check whether we helped them, just like in the prisoner’s dilemma against a copy we don’t have to be able to check whether the other copy actually cooperated. It’s enough to know, prior to making one’s own decision, that the copy reasons similarly about these types of problems.

Comment by caspar42 on Conservation of Expected Evidence · 2017-05-29T13:30:03.203Z · score: 1 (1 votes) · LW · GW

Closely related is the law of total expectation: https://en.wikipedia.org/wiki/Law_of_total_expectation

It states that E[E[X|Y]]=E[X].

Comment by caspar42 on Invitation to comment on a draft on multiverse-wide cooperation via alternatives to causal decision theory (FDT/UDT/EDT/...) · 2017-05-29T13:14:30.308Z · score: 0 (2 votes) · LW · GW

Then it doesn't work unless you believe in some other theory that postulates the existence of a sufficiently large universe or multiverse, Everett is only one option.

Comment by caspar42 on Are causal decision theorists trying to outsmart conditional probabilities? · 2017-05-18T20:35:53.305Z · score: 1 (1 votes) · LW · GW

Another piece of evidence is this minor error in section 9.2 of Peterson's An Introduction to Decision Theory:

According to causal decision theory, the probability that you have the gene given that you read Section 9.2 is equal to the probability that you have the gene given that you stop at Section 9.1. (That is, the probability is independent of your decision to read this section.) Hence, it would be a mistake to think that your chances of leading a normal life would have been any higher had you stopped reading at Section 9.1.

Comment by caspar42 on Are causal decision theorists trying to outsmart conditional probabilities? · 2017-05-18T13:32:32.936Z · score: 0 (0 votes) · LW · GW

Great, thank you!

Comment by caspar42 on You May Already Be A Sinner · 2017-05-18T12:11:57.808Z · score: 2 (2 votes) · LW · GW

In chapter 0.6 of his book Evidence, Decision and Causality, Arif Ahmed also argues that Calvinist predestination is like Newcomb's problem.

Comment by caspar42 on Hofstadter's Superrationality · 2017-02-01T10:27:31.891Z · score: 0 (0 votes) · LW · GW

At the time he wrote it, the correct choice would have been to defect, because as Hofstadter noted, none of his friends (as brilliant as they were) took anything like that reflexive line of thought. If it were done now, among a group of Less Wrong veterans, I might be convinced to cooperate.

I would advocate the opposite: Imagine you have never thought about Newcomb-like scenarios before. Therefore, you also don't know how others would act in such problems. Now, you come up with this interesting line of thought about determining the others' choices or correlating with them. Because you are the only data point, your decision should give you a lot of evidence about what others might do, i.e. about whether they will come up with the idea at all and behave in abidance with it.

Now, contrast this with playing the game today. You may have already read studies showing that most philosophers use CDT, that most people one-box in Newcomb's problem, that LWers tend to cooperate. If anything, your decision now gives you less information about what the others will do.

Comment by caspar42 on Lost Purposes · 2016-09-03T10:26:58.830Z · score: 0 (0 votes) · LW · GW

Many examples are given in the C2 Wiki article on Old Rules With Forgotten Reasons.

Comment by caspar42 on Desired articles on AI risk? · 2016-05-06T22:50:46.219Z · score: 1 (1 votes) · LW · GW

Regarding impossibility results, there is now also Brian Tomasik's Three Types of Negative Utilitarianism.

There are also these two attempted formalizations of notions of welfare:

Comment by caspar42 on Top 9+2 myths about AI risk · 2015-07-09T08:06:34.901Z · score: 1 (1 votes) · LW · GW

Certainly a good compilation! It might be even more useful, though, if it contained references to research papers, Bostrom's superintelligence etc., where the arguments are discussed in full detail.

Comment by caspar42 on Top 9+2 myths about AI risk · 2015-07-09T08:01:38.956Z · score: 1 (1 votes) · LW · GW

Is there a write-up of your objections anywhere?

Comment by caspar42 on Two-boxing, smoking and chewing gum in Medical Newcomb problems · 2015-06-30T20:17:09.835Z · score: 0 (0 votes) · LW · GW

The role that would normally be played by simulation is here played by a big evidential study of what people with different genes do. This is why it matters whether the people in the study are good decision-makers or not - only when the people in the study are in a position similar to my own do they fulfill this simulation-like role.

Yes, the idea is that they are sufficiently similar to you so that the study can be applied (but also sufficiently different to make it counter-intuitive to say it's a simulation). The subjects of the study may be told that there already exists a study, so that their situation is equivalent to yours. It's meant to be similar to the medical Newcomb problems in that regard.

I briefly considered the idea that TDT would see the study as a simulation, but discarded the possibility, because in that case the studies in classic medical Newcomb problems could also be seen as simulations of the agent to some degree. The "abstract computation that an agent implements" is a bit vague, anyway, I assume, but if one were willing to go this far, is it possible that TDT conflates with EDT?

Comment by caspar42 on Two-boxing, smoking and chewing gum in Medical Newcomb problems · 2015-06-30T16:08:48.080Z · score: 1 (1 votes) · LW · GW

But what about when people learn about the setup of this particular problem? Does the correlation between having the one-boxing gene and inclining toward one-boxing still hold?

Yes, it should also hold in this case. Knowing about the study could be part of the problem and the subjects of the initial study could be lied to about a study. The idea of the "genetic Newcomb problem" is that the two-boxing gene is less intuitive than CGTA and that its workings are mysterious. It could make you be sure that you have or don't have the gene. It could make be comfortable with decision theories whose names start with 'C', interpret genetical Newcomb problem studies in a certain way etc. The only thing that we know is that is causes us to two-box, in the end. For CGTA, on the other hand, we have a very strong intuition that it causes a "tickle" or so that could be easily overridden by us knowing about the first study (which correlates chewing gum with throat abscesses). It could not possibly influence what we think about CDT vs. EDT etc.! But this intuition is not part of the original description of the problem.

Comment by caspar42 on Two-boxing, smoking and chewing gum in Medical Newcomb problems · 2015-06-30T15:45:11.891Z · score: 1 (1 votes) · LW · GW

Do you see that assuming Omega worked the way I described, then the number and distribution of boxes containing $1M is exactly the same in the two multiverses, therefore the second multiverse is better?

Yes, I think I understand that now. But in your version the two-boxing gene practically does not cause the $1M to be in box B, because Omega mostly looks at random other genes. Would that even be a Newcomblike problem?

I think this is what makes your version of GNP different from MNP, and why we have different intuitions about the two cases. If there is someone or something who looked the most common gene correlated with two-boxing (because it was the most common gene correlated with two-boxing, rather than due to a coincidence), then by changing whether you two-box, you can change whether other UDT agents two-box, and hence which gene is the most common gene correlated with two-boxing, and hence which gene Omega looked at, and hence who gets $1M in box B.

In EY's chewing gum MNP, it seems like CGTA causes both the throat abscess and influences people to chew gum. (See p.67 of the TDT paper ) (It gets much more complicated, if evolution has only produced a correlation between CGTA and another chewing gum gene.) The CGTA gene is always read, copied into RNA etc., ultimately leading to throat abscesses. (The rest of the DNA is used, too, but only determines the size of your nose etc.) In the GNP, the two-boxing gene is always read by Omega and translated into a number of dollars under box B. (Omega can look at the rest of the DNA, too, but does not care.) I don't get the difference, yet, unfortunately.

In MNP, there is no corresponding process searching for genes correlated with gum chewing, so you can't try to influence that process by choosing to not chew gum.

I don't understand UDT, yet, but it seems to me that in the chewing gum MNP, you could not chew gum, thereby changing whether other UDT agents chew gum, and hence whether UDT agents' genes contain CGTA. Unless you know that CGTA has no impact on how you ultimately resolve this problem, which is not stated in the problem description and which would make EDT also chew gum.