![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
gusl31 Dec 2002
2pm I go to Porto de Galinhas, not knowing where I'll be sleeping.
4pm Rudy calls, and we meet up.
5pm Get lunch. Walk all along the beach. See many tourists. Play a blues riff with a cavaquinho (kinda like a ukulele) with some hippies. The city was very full: I must have seen like 15 Asians (that's my new handy heuristic for estimating crowds in Brazil).
7pm Watch a very entertaining accordion player / singer / dance caller.
11pm Play porrinha. Figured out a (in retrospect) trivial result about porrinha with rational players. If you assume that it is common knowledge that all players are rational (the strongest possible assumption, I guess), then each player's best guess is based only on the previous player's guess and his own hand (greedy analysis).
1 Jan 2003 - 12:05am NEW YEAR! Hugs going all around, which was nice. It's interesting to be hugged by someone you've never spoken with. Saw the clouds being lit up by the fireworks on the next beach. It was beautiful.
1:00am Saw my cousins Renata, Maria, Gabriela; and Felipe and Pedro, who grew up in California and who give people a chance to practice their English. Then I went to talk to the older generation.
Extremely geeky conversation with Roberto (my physicist cousin who lives in San Francisco, father of the two boys above): he's still working with models for stock-picking using AI techniques, etc, but recently got back to working on physics again: he's doing very interesting work with Constantino Tsallis (a 3000 hit man!), and he's given a talk about the work at the Santa Fe Institute.
I impressed him with the paradox of mathematical information (aka the paradox of "in logic class you learn nothing"). We also talked about Hofstadter, and other popular science personalities.
I told him I would love to do this kind of research. He said it's looking difficult at the moment, but we kept talking about the stuff, so who knows what will happen when I follow up. It's so cool to know another geek in the family, even if we're that far (1/16 blood).
2:30am Check out the rock band, who played a few songs from Abbey Road remarkably well.
2:40am Talk to Roberto some more. He explains Tsallis's ideas in more technical terms. Interesting: Gaussian distributions may actually be less "natural" than longer-tailed distributions in some cases. (The Gaussian is the maximum entropy distribution (i.e. "natural") under the classical (i.e. less general) assumptions). The consequence is that when those assumptions don't hold, supposedly "rare" events happen more often than you would expect. Very cool stuff. They say the consequences show up in biology, finance, and a lot of "complex" fields (even if skepticism is called for whenever you hear hyped claims about "complex systems", "non-linear dynamics" or "chaos").
3:00am Rudy and I go walk towards the Skol place, then give up and get the car, and then we give up because of the traffic, so we park halfway and walk along beach. The Skol place was really sucked and was really expensive. So we walked back.
3:50am bought an Orange Pitu. Enjoyed hard rock like I hadn't in a long time.
4:15am felt more inebriated than I remember ever being, after having only half of a cup.
4:30am Rudy sees me trying to sleep with my head on a plastic bottle, and says we should go home.
5:00am We get home.
~6:00am I get to sleep on a very tight couch.
9am Wake up very sweaty, by people arguing.
10am offered a sleeping bag.
12noon wake up for good
2:30pm arrive home
2:40pm get a fix of the 'Net
2pm I go to Porto de Galinhas, not knowing where I'll be sleeping.
4pm Rudy calls, and we meet up.
5pm Get lunch. Walk all along the beach. See many tourists. Play a blues riff with a cavaquinho (kinda like a ukulele) with some hippies. The city was very full: I must have seen like 15 Asians (that's my new handy heuristic for estimating crowds in Brazil).
7pm Watch a very entertaining accordion player / singer / dance caller.
11pm Play porrinha. Figured out a (in retrospect) trivial result about porrinha with rational players. If you assume that it is common knowledge that all players are rational (the strongest possible assumption, I guess), then each player's best guess is based only on the previous player's guess and his own hand (greedy analysis).
1 Jan 2003 - 12:05am NEW YEAR! Hugs going all around, which was nice. It's interesting to be hugged by someone you've never spoken with. Saw the clouds being lit up by the fireworks on the next beach. It was beautiful.
1:00am Saw my cousins Renata, Maria, Gabriela; and Felipe and Pedro, who grew up in California and who give people a chance to practice their English. Then I went to talk to the older generation.
Extremely geeky conversation with Roberto (my physicist cousin who lives in San Francisco, father of the two boys above): he's still working with models for stock-picking using AI techniques, etc, but recently got back to working on physics again: he's doing very interesting work with Constantino Tsallis (a 3000 hit man!), and he's given a talk about the work at the Santa Fe Institute.
I impressed him with the paradox of mathematical information (aka the paradox of "in logic class you learn nothing"). We also talked about Hofstadter, and other popular science personalities.
I told him I would love to do this kind of research. He said it's looking difficult at the moment, but we kept talking about the stuff, so who knows what will happen when I follow up. It's so cool to know another geek in the family, even if we're that far (1/16 blood).
2:30am Check out the rock band, who played a few songs from Abbey Road remarkably well.
2:40am Talk to Roberto some more. He explains Tsallis's ideas in more technical terms. Interesting: Gaussian distributions may actually be less "natural" than longer-tailed distributions in some cases. (The Gaussian is the maximum entropy distribution (i.e. "natural") under the classical (i.e. less general) assumptions). The consequence is that when those assumptions don't hold, supposedly "rare" events happen more often than you would expect. Very cool stuff. They say the consequences show up in biology, finance, and a lot of "complex" fields (even if skepticism is called for whenever you hear hyped claims about "complex systems", "non-linear dynamics" or "chaos").
3:00am Rudy and I go walk towards the Skol place, then give up and get the car, and then we give up because of the traffic, so we park halfway and walk along beach. The Skol place was really sucked and was really expensive. So we walked back.
3:50am bought an Orange Pitu. Enjoyed hard rock like I hadn't in a long time.
4:15am felt more inebriated than I remember ever being, after having only half of a cup.
4:30am Rudy sees me trying to sleep with my head on a plastic bottle, and says we should go home.
5:00am We get home.
~6:00am I get to sleep on a very tight couch.
9am Wake up very sweaty, by people arguing.
10am offered a sleeping bag.
12noon wake up for good
2:30pm arrive home
2:40pm get a fix of the 'Net
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
(no subject)
Date: 2003-01-01 09:27 pm (UTC)The site you link too is quite long. Is there a concise definition of the paradox of mathematical information?
I apologise in advance for my ignorance.
(no subject)
Date: 2003-01-02 07:05 am (UTC)Basically it can be said that one learns nothing in a math class, since all theorems inevitably follow from the axioms. In other words, if you know the axioms, then you know all the theorems. If you were to measure how much new information a theorem gives you, you would always get 0, since it already followed from the axioms.
However, this is absurd because it ignores all the computations that may be required to spit out a theorem. We need a new interpretation of "know" which takes into account the problems of deduction and knowledge retrieval. (If you have my phone number in you Palm pilot, then is it true that you know my phone number? What if you had to do a complicated, but finite calculation?)
The point is that classical information theory is inadequate for talking about axiomatic systems. Information theory was originally only applied to raw data. Shannon once said that information theory had nothing to do with meaning or truth. So these guys at the link are working on a "semantic information theory", so that they can talk about information in mathematics (I think that's one goal, anyway).
Is this clear?
asians in brazil
Date: 2003-01-02 07:31 am (UTC)> for estimating crowds in Brazil).
i thought sao paolo had something close to the largest japanese population outside of japan, no?
i realize you don't live in sao paolo, but perhaps you should specify for which areas of brazil or for which sub-sets of asians the asianometer theory works :p
- laura (http://www.lauraelgin.com)
Re: asians in brazil
Date: 2003-01-02 07:33 am (UTC)- laura (http://www.lauraelgin.com)
(no subject)
Date: 2003-01-02 09:50 am (UTC)As for the explanation, it's all clear to the third par when my ignorance of classical information theory means I'm going to have to go looking for a concise definition of what that's all about.
I'm assuming that semantic information theory is in some way an extension of the work of Frege and Russell which I do know a little bit about. This is off the subject of mathematics but I've sometimes thought that although their attempts to replace vague ordinary language with precise logical language were admirable in that they led to some extraordinary work, these attempts were unnecessary in themselves. This is because the vagueness in everyday language can be pointed out and combatted using the very same tool of vague, everyday language.
I do recognise, however, that you could potentially end up needing an infinite series of clarifications with ordinary language. These would be aimed at clarifying the obfuscations introduced in earlier iterations of the attempt to clarify with ordinary language obfuscations that it had already given rise to.
Having said that, you could ask just how much clarity in ordinary language you need for any practical task. Just as you might only need to work out a mathematical question to a certain number of significant figures or decimal points, so you might only need in practice to answer a question embodied in ordinary language to a certain degree of clarity.
Back on topic, you could make the same argument that you use with maths in the area of economics. Even the attempt to replace the classical economic man with a more evolutionary approach to the subject could again be considered empty of a posteriori knowledge if it just starts with game theory based upon axioms.
Only economics which starts with empirical results from experiment and observation could be considered to be a posteriori, even if you could argue that the complex architecture of theories that have won Nobel prizes, swayed economics ministries and made millions rich and poor in turn amounts to a form of synthetic knowledge.
I hope I've not got terribly confused about Kant. It's so hard not to be as a mere beginner who can't be bothered to read his tortuous prose nor that of those who write about him.
I've read that there's something called post-autistic economics stirring in France but I don't know if they're addressing this question. Again, lack of time and patience for wading through pages of websites leaves me grasping for lifebelts of knowledge in my sea of ignorance.
(no subject)
Date: 2003-01-02 09:55 am (UTC)Re: asians in brazil
Date: 2003-01-02 11:57 am (UTC)The Asianometer probably works for the part of Brazil I'm concerned with, since I hardly ever leave Recife.
I guess it can be pretty inaccurate, due to biases (i.e. beaches tend to have more tourists, etc.)
Happy new year to you too.
(no subject)
Date: 2003-01-02 12:34 pm (UTC)I am totally into formalizing language, and this is one my goals for AI. In a way it's backwards NLP: make humans speak logically instead of having computers try to understand ambiguous, vague, illogical statements. Rhetoric can be a nice art, but it should be left to the actors and poets.
I am skeptical of most infinity claims. The human mind is finite. Human language is finite. Can you challenge me with a specific issue?
Re: asians in brazil
Date: 2003-01-02 09:33 pm (UTC)- laura (http://www.lauraelgin.com)