Here is our monthly place to discuss Overcoming Bias topics that have not appeared in recent posts.
Why are fathers present at the birth of children? How in 20 years did something unprecedented become almost mandatory? Particularly with no evidence that the practice is of any benefit
“A top obstetrician on why men should NEVER be at the birth of their child”
I would like to know what I should do if I want to advance the state-of-the-art in artificial intelligence. That is, what formal education should I get, what works should I read relative to the other steps, what groups and projects I should join, what hobbies I should take up, where I should live, etc.
And let’s say I also hold a bachelor’s degree in mechanical engineering and have been reading popular-press books related to AI: Godel, Escher, Bach; The Emporer’s New Mind; Dennett’s and Yudkowsky’s stuff, etc.
That article is dripping with confirmation bias and generalization (anecdotes as data).
That doesn’t mean he’s wrong, but more evidence than a single doctor’s gut feeling is needed.
Is there an explination for these seemingly contraditory stereotypes:
Daughters turn into their mothers
Husbands love their wives
Husbands strongly dislike their mother-in-laws
Wouldn’t you think if one is good two is better, or does the marginal return become negative at some point shortly after one.
meta issue: I’ve noticed sometimes that it looks like new comments get inserted “in between” older ones.
That is, sometimes I’ll look at the comments for some posting, then later look again and the last comment may be the same, but it seems like new stuff gets inserted in between some of the older ones. Is it just me misremembering or is this some oddity in the way commenting works here?
Could be comments being held back for manual approval by a spam filter. I’ve never seen it (on this blog), though.
Nick: you’ve never seen spam on this blog because the site’s captchas are capable of generating characters that can’t be consistently identified by any existing algorithm, despite Eliezer’s claim to understand the algorithm by which his brain identifies objects. 😉
Okay, I’m so banned.
>That article is dripping with confirmation bias and generalization (anecdotes as data).
>That doesn’t mean he’s wrong, but more evidence than a single doctor’s gut feeling is needed
I agree. The other side (men should be in the delivery room) does not seem to have provided even this low level of evidence. Not that this issue is major thing but it might indicate how some of our medical actions are not based on evidence but on social pressures.
Would anybody else be interested in another OB meetup in the Bay Area?
the site’s captchas are capable of generating characters that can’t be consistently identified by any existing algorithm
Humans can’t consistently identify them, either – just sufficiently better than existing programs to permit us to make good judgments when we try to distinguish them.
Do not fall into the error of thinking that humans can do things which algorithms cannot.
I am interested in another OB Bay Area meetup… why don’t you create a meetup group?
CREATE A FORUM FOR OVERCOMINGBIAS.COM
I have the impression that most reader comments get buried here. Solution:
1. Create a main forum with exactly one thread for every post on this site where people can reply to the thread.
2. Create another forum where people can start their own threads.
This would be a way to get this thing organized.
I agree, but the burden of proof is different when denying someone the ability to do something they want to do. There would have to be a respectable amount of evidence to show it’s a significant risk factor to have husbands present.
I was present as the primary support person when both my sons were delivered, both because I wanted to be, and my wife wanted me to be there. They would have had to come up with some very compelling evidence to eject me from the delivery room.
Had I not been allowed to be there (and assuming we knew this was going to be the case before we showed up there) we would have very likely attempted a home delivery, at least in the first case. The second would have obviously been contingent on the outcome of the first.
I for one am not familiar with the “proven fact” that it is impossible to have a relaxed, quiet mind when next to a person hyped up on adrenaline, and I would like very much to see some references for that claim.
It is not clear to me that the things the ob complains about hold in most cases – or even in any of them. Data are very much needed, and he cannot or will not supply them.
I have the impression that most reader comments get buried here.
WordPress counteracts this problem by giving each blog entry its own RSS feed. Provided the user is already in the habit of checking an RSS reader regularly, subscribing to an entry’s feed eliminates the need to remember to return to the blog regularly to check for a new comment.
A typical example is here where it says, “You can follow any responses to this entry through the RSS 2.0 feed.”
What do OB readers think the next “Age” is going to be, i.e. Industrial Age -> Information Age -> ?
I started thinking about this after reading the post about the self-replicating factory. I always thought Biotech would be next but maybe it will be molecular precision manufacture, ala Neal Stephenson’s Diamond Age.
Silas: I was thinking about those questions, too, you seem to have created the most comprehensive list I’ve seen 🙂 Eliezer advised that a) we should get rich and invest into IT b) learn computer programming. First one… well… under construction, and for point b): but then work on what? I don’t think working at a big company with a lot of lost purposes and coding some dull stuff as the nth drone of hundred programmers makes the world that much a better place… So what do you think?
(Yet another example for a “college student who wants to help but got lost”: a year till an MsC in electrical engineering with a lot of NN, Bayes-nets, math etc stuff, some programming skills, reading OB, GEB etc, with a desire to save the world :P)
Does this blog have an RSS comments feed for each entry? If so, it would be helpful to make this (more) accessible.
Here are a few:
1. Can anyone point to a (non-wikipedia)resource which has a nice crisp taxonomy of specific biases, their hypothesized geneses and evolutionary pedigree, and specific structural analytic techniques that can help us mitigate specific biases. Is there such a thing out there?
2. Why do “heuristics” and “biases” seem to be all semantically kluged together in various discussions? Are they two fundamentally different things? Are heuristics simply evolved cognitive shortcuts resulting from environmental forces, and not particulary value laden, while biases are a result of the heuristic itself and entirely value-laden? Or is that totally off-angle?
3. How do I become adept at Bayesian analysis? What truth model do you think Bayesian rationalism best supports?
Ian C. –
I’d vote for the next “age” to be the intelligence age. Whether it’s human augmentation (chemical, genetic, bionics), or independent AI, I think that’s the next (last?) frontier.
Latinus: glad to see someone else is in the same position as me. Like you, I’m “working on” 1), and as for 2), I can program, yeah, but that’s mostly confined to “do computation and output the data as text”. I’ve been trying to expand this to cover richer interfaces by learning Visual C#, but I, er, keep putting it off.
Another thing I was planning to do was work on building an android and adding in all the open-source robotic stuff. And yet another would be downloading the neat “genetic algorithms” from the academic sites linked from youtube. I’ve seen some pretty neat stuff, like the four-legged robotic spider that “comes to life”, models its world, and figures out how to walk around; and the organism simulators that figure out how to move around in a virtual environment.
I won’t link youtube, though, because that raises the probability of this message being marked as spam.
So … that’s why I’m asking for guidance like you 😉
(email address is email@example.com, can’t put it in the URL field or otherwise make it show)
AI as a discipline has currently lost its way, it has fragmented into a million different potential useful techniques/architectures.
You need a strong theoretical sense in order to pick one program of research over another. If you are happy to create a cool specific purpose machine (e.g. car driving) then Stanford or MIT might be good. If you want something a bit more human level, you have to find your own path (some of the few general systems in academia are here at AIXI or Goedel machines) or go a bit out of academia.
You could try the agi list run by the agiri, or ai-philosophy on yahoo, for debate about AI.
Other things useful to know are about the No Free Lunch Theorems of search. Reading Machine Learning, will get you up to speed on induction. Although I don’t think induction should be the bedrock of an AI architecture, it is very important.
How about whether and how patents might stimulate human capital development?
Michael Martin: As for patents (and copyright), see the book Against Intellectual Monopoly (freely available here , and here is a good review with the most important points, they sound reasonable.)
Silas: does the spam filter notice youtube video ID-s too? (I haven’t tried it yet…) But I think it would be easier to use virtual worlds first, coping with noisy sensors and slow actuators seem to be a typical narrow AI problem. (I think Novamente does something like that, intelligent agents for Second Life).
Will Pearson: thanks for the advice, that’s at least one more useful book to read 🙂
Latanius: Good point, and I probably would have thought of it if I didn’t have additional barriers besides that (couldn’t access youtube at the time).
Here is the real-environment robot I mentioned, very cool, and has a link to the university site: yt video ID ehno85yI-sA
And here is a survey of virtual animals to simulate evolutionary algorithms: yt video ID AHBNbcDpjeU
I do not want to subscribe to an RSS feed for each topic that I want to follow. Can you imagine what your RSS reader would look like after a year of doing that for all your websites? More to the point, what a freakin’ pain in the neck!
Having separate websites on separate things is a form of organization. Conglomerating everything you read into a single RSS reader would be disorganization. Then you would need to create filters for your RSS reader, to separate out the different feeds into the websites they originally came from. That would be a lot of work for something equivalent to just checking the original websites.
I would like to come to Overcoming Bias, and have it work like a forum. That would be nice.
Let’s create a “tip” prize for whoever puts such a forum together.
I’ll find a way to pay the founder of the forum $10.
How much will you find a way to pay them, Phil?
(The forum could be founded by anyone, doesn’t have to be the actual OB admins.
I do not wish to take the momentum out of Phil Goetz’s and HA’s forum ambitions but it might interest Eliezer to know that Sonic Dot Net has started giving away free WordPress blogs to every customer of theirs. He probably got an email two weeks ago from the CEO of Sonic Dot Net announcing this. WordPress is a GPLed collection of PHP, HTML and CSS files that sits in the Sonic customer’s public_html directory. I hear that there is no delight in hacking on this code, but the (web) interface to WordPress is pleasant and very, very competently engineered. I am a (delighted) Sonic Dot Net customer and got me a WordPress blog now.
You’ve heard of BarCamp? Well, there are WordPressCamps (name?) in San Francisco and other cities. How is that for a passionate community (of blog owners and such)?
Per-entry or — to use the official WordPress term — per-article RSS feeds are probably specific to WordPress. This is a Typepad blog. Typepad seems to be a proprietary fork of Moveable Type, which is the second-most important (popular) blog-software ecosystem, the most important (popular) ecosystem being WordPress.
The name is not WordPressCamp: it is WordCamp.
Richard, minor correction: TypePad isn’t a proprietary fork; it’s a service operated by the same company that makes Movable Type.
Well, maybe I should have chose Movable Type rather than WordPress for my own blog then.
Roland, Phil, HA: I created a forum. Posts will be automatically mirrored. (HA, you can PayPal $10 to firstname.lastname@example.org .)
Last step: put a link to the forum in the sidebar.
Nick, got it. Community, feel free to adjust my reputation accordingly.
A semi silly thought I just had:
Q: Why do Bayesians consider it so important to be able to quantify one’s ignorance?
A: “If it has stats, you can kill it.”
Bayes theorem gets mainstream (or at least 2nd standard deviation) attention:
Also hints at how applied math phd’s are bringing back generalist scientists.
This Paul Graham essay is interesting, but I took particular notice of these two paragraphs:
“Innocence is also open-mindedness. We want kids to be innocent so they can continue to learn. Paradoxical as it sounds, there are some kinds of knowledge that get in the way of other kinds of knowledge. If you’re going to learn that the world is a brutal place full of people trying to take advantage of one another, you’re better off learning it last. Otherwise you won’t bother learning much more.
Very smart adults often seem unusually innocent, and I don’t think this is a coincidence. I think they’ve deliberately avoided learning about certain things. Certainly I do. I used to think I wanted to know everything. Now I know I don’t.”
I think most of the issues he raises have been touched on here at some point or another; the quoted portion made me think of Robin, who seems innocent in that way despite his stated desire to know the whole truth.
The scale of true human intelligence is measured within our own creativity. As smart as we are, we have to invent to be considered intelligent (ideas or otherwise). The very existence of overminds and Jupiter brains are creations of our own intellect. In fact, as a purely fictional character, an overmind is a particularly impossible character to write for because there is no possible way to explain its motives from a human point of view, and given that explanation is the key to communication, there is no possible way to consider an overmind magnificently intelligent without going past your own logical understanding of life. The purest form of intellect comes from a few words: Do unto others; Life is suffering; Logic is little tweeting bird chirping in meadow. Logic is wreath of pretty flowers that smell bad.
I hope they stay this time.
… be a charity angel.