We've just had a nice intro to causality from Eliezer. What if we have some variable nodes that we think might be relevant, but which we cannot observe? Judea Pearl still has a lot to say about that: as long as you can observe the right subset of variables to d-separate the source and target node, you can identify causal effects. What if you can't observe the right subset? Hope would seem to be lost, but if you aren't worried about determining the exact strength of the causal effect, but instead just determining if the strength is non-zero, you still have a chance.

You may remember a bunch of newspaper articles a few years ago claiming Obesity is contagious. That study's conclusion was based on modeling assumptions (review), but Shalizi and Thomas considered the causal picture and found it to be bleak: exactly identifying the causal effect is problematic because of latent homophily. I.e. humans can have many hidden attributes that affect who we become friends with and whether we have a tendency to become obese. Measuring all these human factors is hopeless and interventional experiments (where we randomly force some people to become obese and then observe their friends) is unethical.

I'm happy to add a little to this conversation with my recent paper showing that latent homophily does not prevent us from finding causal relationships in social networks. I show that even if you can't measure all the relevant hidden attributes describing humans, you can lower bound the strength of causal effects. We're mostly interested in whether the effects exist, and only secondarily how strong they might be. Of course we still have to make some structural assumptions (this is the essence of Pearl's work: it allows us to infer causality given general, intuitive, structural assumptions among variables). E.g., there is no external dynamic force (like a deity) that causes Alice and Bob to become friends and then causes them to become obese at the same time. Less extravagantly, how about a systematic, widespread change in the consistency of diets? So, although the conclusion of the paper is: latent homophily alone cannot explain correlations in obesity, we still can't say for sure that your friends make you fat. This is based on older work which considered a little more general setting about tests to rule out broader classes of hidden variable models, like the Bell inequalities.

I'm writing a paper on how recent advances in the neuroscience of decision-making should affect our understanding of welfare economics, and consequentially impact an approach to AI safety. I'll be seeking peer-review for the paper, and if accepted it would be my first academic publication!

My current drafts are private, so I can't link at the moment. But if there is an interest, I'll be sure to post a link once I make my work public.

I'm working on a non-Elcenia novel, Goldmage. I'm stuck in the middle of chapter five at the moment and haven't quite worked up to skipping the sticky scene and doing what's next instead.

I'm currently researching startup concepts surrounding two main themes - big data analysis/visualization and scientific research. I have a plan for making this happen, and at the current stage I'm setting as many meetings as possible with people who know about these topics. The goal is to map out how science works - where the money comes from, who does what, how labor is divided, what the problems are - and then start isolating big problems in the space that might be solved through data analysis or visualization. After that, I test and develop a business model hypothesis via Steve Blank's startup development process (as described in the Startup Owner's Manual).

But anyway, back to this month: I'm setting as many meetings as possible with scientific researchers, people who run labs, R&D managers, people in the NSF or other organizations, and other relevant individuals. So if any of you fall into these categories I'd love to talk to you! Private message me.

Having temporarily put my bigger business idea (quit-job scale) on hold, working on a smaller one (side-project scale) to gain skills with web development and running a site. It's oriented towards people who like reviewing books, so should be interesting to a number of people here. More details to follow... ETA ~ 2 months or less.

I just finished my NaNoWriMo novel, Judge on a Boat (latest revision kept here), last month in November, and this month I'm going through the process of fixing it up and improving it. I described it on LessWrong yesterday.

Why this project? Well, I've been lurking on Less Wrong (and before that, Overcoming Bias) for years, and yet I recently realized that I've not been very rational in actual practice. So I decided to write a novel about rationality and moral philosophy, just to make sure that I managed to actually understand the topics well enough to put them in my own words. Hopefully the attempt to explain them to a lay audience will help my own understanding.

I'd like to get some help from others in the LW community, since I suspect the novel is not very well-written, and I need some ideas on how to improve it. Why should anyone help me? Well: two of the best recent rationalist fiction that I know of are Alicorn's Luminosity and EY's HPMoR. I am nowhere near those levels (for one, their characters are not flat). The only advantage I have is that my novel has (in current law, anyway) a slightly higher chance of being published, unless J.K.Rowling suddenly has an aneurysm and gives the copyright to the public domain, or if suddenly everyone listens to rms and start repealing copyright laws internationally: the novel is original and won't get sued into oblivion if published.

My goals are... a bit iffy. I imagine publishing this in actual real-world physical book form, because those things are easier to give as gifts and might help raise the sanity waterline (badly needed in my family, and least they read books). But with the current level of quality I suspect I have about a snowball's chance of passing unscathed through the sun.

Alternatively: how about an open-source novel? I could put it up into a CC-BY-SA and try to actively recruit people to help improve it, try to leverage the community, but that probably will make it difficult to publish physically, as legally speaking (IANAL) that would require contacting all the copyright owners. Maybe a fiduciary agreement a la FSF-Europe, but I know of no big, trustable entity that would act as a fiduciary for fiction.

I am writing a blog about July Rationality Minicamp and the previous week which I spent in SF Bay Area. But I am writing it really slowly. Here are the written parts: 1 2 3 4 5, but only the last two of them are about the Minicamp. I write them in both Slovak and English, and the idea is that I will later use them for trying to recruit some rationalists in Slovakia.

Meta: Now I looked to my comments in previous threads. The game I was working on in April is still not ready. I had some problems with multithreading that I did not know how to solve. I solved them a few months later, but I lost the momentum. But I still want to finish that game. I have finished the high-school textbook on informatics I was writing in February, but we had a change of government, and the new government postponed the textbook projects, because apparently the education is not a priority.

I'm helping to write the 2013 MIT Mystery Hunt. I have been spending more time editing and less time authoring puzzles than I'd like; that's probably fine, since my skill at the former noticeably exceeds my skill at the latter.

Things interesting to LW:

I'm reading Value Focused Thinking and plan to post my notes on LW. It's a lighter book than past offerings, and so I hope it will be more like a summary than a description. Why this book? I recently attended a seminar by the author and was impressed; I think there should be more decision analysis posts on LW, and more posts about discovering values, and this looks to be a useful book for both of those ends.

I've been incorporating stuff from the 3rd season of MLP into my fanfiction (among other things, I got a far better introduction and title from the first episodes, and future ones are adding more bit by bit). I'm still way farther than I want to be from starting to make it public, and have not been putting as much time into it as I'd like. I think that will change once I finish up school projects (which needs to happen this week).

I have roughly >doubled the size of my (Japanese) anime/manga biracial character database to 273 characters by exploiting some databases I was ignorant of before, like AniDB & VNDB: http://www.gwern.net/hafu#list

This large expansion adds substantially to the statistical quality of any conclusions I'll draw. So far the only new thing I've noticed is that there is now a glaring absence of Korean characters in my database, despite the long intimate history of Korea & Japan, and the fact that without even looking or compiling a similar database, I have found ~4 half-Japanese characters in Korean works.

I also split out my iodine info and meta-analysis as a separate page: http://www.gwern.net/Iodine

I've done some further work on my dual n-back meta-analysis as well: regressing against time allotted for IQ test (no effect) and hopefully soon against the type of IQ test (hypothesis: BOMAT tests inflate IQ improvement estimates).

I've finally started thinking about logical uncertainty again. So, maybe decision theory blog post forthcoming?

Of course one can simply say that the expected value of betting on logical propositions is higher than that of not betting. But not only is this weaker and possibly more complicated than simply proving an analogy of one of various derivations of probability, it's a lot easier to say the words than to show it for an algorithm without circularity.

I have been a bad composer, and a bad blogger. BUT recently I have been better at self-promotion (having put myself forward for more opportunities than I normally would), a better teacher, and a better pianist (I'm making technical breakthroughs and getting back into a good practice routine).

I know some of the reasons that I'm not composing and blogging as well as I'd like, and I have plans to deal with them. I also launched my soundcloud after putting it off for quite a while - file under self-promotion too, I guess.

I'm working on adding elements to a report at work that does data visualization on a large scale (the data set is about 1 million data points; it's really not all that impressive of a subject matter, but I can't be terribly specific). The report has all of the "easy" elements I need in it, but now I'm trying to add in the harder elements. My ultimate end goal would be to add in the more complicated data along with system parameters, so I can get a handle on how parameter changes affect the output. I'd love to see Bayes nets and the like make a triumphant entrance at some point. But near as I can tell, I'd be the local expert on all of that, and anything I know about that subject matter I mostly picked up from here.

• I am currently learning to use the Neo2 keyboard layout. It is slow but I am positive that it will pay off in the long run.
• I work to be happier. Again a very slow process. I start by enforcing a sleep schedule via melatonin, seeing as being sleepy ruins everything.