Gustavo Lacerda

writing for a specialized audience

Wed, 23 Dec 2009 17:16:03 GMT

Right now I'm reading William Cohen's book "A Computer Scientist's guide to Cell Biology", and I find the delivery to be very efficient (though I have little to compare it with), probably because he takes an informational perspective, and isn't shy about using CS concepts and terminology.

You know the "X for dummies" collection? I'd love to see some "X for geeks" series. It could be specialized into "X for mathematicians", "X for Computer Scientists", "X for type theory geeks", etc.

According to Sussman, the legacy of Computer Science is its formal language:

<< Computer Science is not a science, and its ultimate significance has little to do with computers. The computer revolution is a revolution in the way we think and in the way we express what we think. >>

I suspect that, when most scientists speak this "language", we will see greater understanding across disciplines. This is already happening.

---

Tangentially, I'd like to see a book on how to cope with "bad" programming languages, all the while being a hygienist. e.g. tricks for emulating a type system, etc.

Internet censorship in Brazil

Tue, 22 Dec 2009 15:29:25 GMT

Brazil is very far behind when it comes to Internet censorship.

* People have been sued and heavily fined for writing restaurant reviews on their blog.
* YouTube was banned for a week because Daniela Cicarelli complained of a single video that violated her privacy.
* this proposed law puts blogs and social-networking profiles on the same footing as broadcast radio and TV in the weeks prior to an election: you must remain impartial and mention all candidates evenly. i.e. you can't openly support any candidate.

I'd like to see a local EFF -type organization, but I don't.

GMail feature request

Fri, 18 Dec 2009 06:56:37 GMT

When replying to a message in GMail, if you change the subject, GMail creates a whole new thread, fully independent from the original thread. In fact, after you've done it, you can't tell that the message has been replied to at all.

Feature request: when you change the subject, automatically create a link to the new thread, right below the message from which the new thread was spun off.
Plus (optional): also create a backlink, at the top of the new thread.

"A Bayesian World"

Sun, 13 Dec 2009 10:56:25 GMT

This was one of the highlights from the NIPS workshops goodbye banquet:

I was half-expecting the entire room of 200+ to join hands.

I saw them the night before and suggested a pun or two, but they rewrote most of it since then.

on being narrow

Wed, 09 Dec 2009 17:24:58 GMT

NIPS is so broad overwhelming that I need some way to stay sane.

I could have spent Monday's poster session (from 7:30pm to midnight!) just looking at papers about two-sample tests (But that would be a bit extreme).

So as a matter of policy, I've been avoiding papers about:
* reinforcement learning
* vision

... which essentially means that I care about the same problems as statisticians.

It's a blacklist policy, because a whitelist policy would necessarily be more narrow-minded.
I am distinctly noticing a lack of papers about biology.

optimal regularization; losing the Bayesian faith?

Tue, 08 Dec 2009 10:43:55 GMT

Tonight, Percy Liang presented an interesting poster on asymptotically optimal regularization: for a certain class of regularizers, they show how to minimize risk asymptotically.

I was somewhat surprised to learn that this optimal estimation can be interpreted as MAP estimation with a two-level hierarchical prior.

But perhaps my surprise is surprising. I seem to be behaving as if there are no procedures with good frequentist properties that can't be interpreted as Bayesian inference for some prior!

[I wish to leave it ambiguous whether this is sarcastic... as I in fact do not know what to think]

"Sugar: The Bitter Truth"

Mon, 07 Dec 2009 00:39:06 GMT

Best dieting tip ever: wait up to 20 minutes for the satiation to arrive. It's working for me.

This is from "Sugar: The Bitter Truth", by UCSF professor Robert Lustig who said "High Fructose Corn Syrup is Poison", with convincing evidence that it causes metabolic syndrome and the obesity epidemic. http://www.sweetsurprise.com/

"There is something wrong with our biochemical energy feedback system."

"fructose goes way beyond empty calories. It is a poison."

"AFAIAC, this stuff was Japan's revenge for World War II"

HFCS is so cheap that it has found its way into everything: hamburger buns, sauce, ketchup, most loaves of bread.

The Coca-Cola conspiracy: coke has lots of salt, the sodium makes you thirsty; the sugar hides the salt!

Some schools performed the intervention of cutting out coke machines, and it had a significant effect on obesity and type II diabetes.

On a minor point, he made this statistician (and former logician) cringe once or twice (by assuming Gaussianity, interpreting a defeasible argument as a deductive one and declaring "only the contrapositive is transitive") but his central message seems to be sound, and of course very important.

* - ("lustig" means "funny" in German)

a flu by any other name

Fri, 04 Dec 2009 03:20:15 GMT

Mexican flu, swine flu, H1N1, hiney, the 2009 flu pandemic, the "A" flu (romance languages)

What does it about something say if people can't settle on a name?

Source: http://en.wikipedia.org/wiki/2009_flu_pandemic

UBC weather

Thu, 03 Dec 2009 13:18:17 GMT

Does UBC have a weather station with live broadcasting?

I suspect UBC is a degree or two colder than Kitsilano.

Do people ever consider microweather when they make real-estate decisions?

hyper-abstracted R contest

Thu, 03 Dec 2009 05:45:02 GMT


## `*` is the hyper of `+`, `^` is the hyper of `*`
> hyper <- function(fn) function(a,b) Reduce(fn, rep(a,b))

> compose <- function(fn1,fn2) function(x) fn1(fn2(x))

> hyperoperation <- function(n) Reduce(compose,listRep(hyper,n))(`+`)

('rep(obj,n)' and 'listRep(obj,n)' just return a list containing 'obj' n times. I had to invent 'listRep' for technical reasons, namely passing closures to 'rep' returns an error: "object of type 'closure' is not subsettable")


## 'hyper' clearly works:
> multiplication <- hyper(`+`)
> multiplication(2,3)
[1] 6
> exponential <- hyper(hyper(`+`))
> exponential(2,3)
[1] 8
> tetration <- hyper(hyper(hyper(`+`)))
> tetration(2,3) ##2^(2^2)
[1] 16

'compose' also works:


> compose(function(x) x^2, function(x) x+1)(4)
[1] 25

and so does 'hyperoperation', woohoo!


> hyperoperation(1)(2,3)
[1] 6

until R decided to be lame:


> hyperoperation(2)(2,3)
Error: evaluation nested too deeply: infinite recursion / options(expressions=)?
>

But I was impressed that I got this far with R! I suspect I could get further if I learn to use defmacro.

inspi#FF0000 by Scott Aaronson's essay, "Who Can Name the Bigger Number?"

preventing big earthquakes

Mon, 30 Nov 2009 08:54:11 GMT

This document contains what sounds like a crackpot idea: causing small earthquakes in order to prevent large ones. I am not an expert, but their logic seems sensible, so for all I know it might work.

My understanding is that earthquakes happen when two tectonic plates are moving towards each other, and the pressure buildup causes slippage between the (hard) sides of the fault.

I can imagine potential solutions falling into 3 categories:
* release the pressure more gently (e.g. more often), perhaps causing smaller earthquakes. The above idea is an example of this.
* prevent slippage, e.g. by filling faults with hard material
* dampen the slippage, like the above, but using softer material.

collaborative geek humour

Sun, 29 Nov 2009 04:29:45 GMT

Last night I witnessed (and possibly participated in) a great example of my type of geek humour. It was a spontaneous collaborative joke, with 3 or 4 people making suggestions.

A modern couple gets married, and decides they will come up with an entirely new surname. They decide that this will be done during the ceremony, using the input of the guests present. This is done by ______.

To avoid extreme ridiculousness in their new name, they introduce a constraint, namely ______.

I can't remember the exact suggestions... but I'm curious to see what my LJ readers come up with.

lexical differences between PT vs BR Portuguese

Sat, 28 Nov 2009 00:03:05 GMT

I have long said that the differences between PT and BR Portuguese are very analogous to UK vs US English. Sometimes I reconsider this. For example, if you're a Wikipedian, there's some annoying differences between the Portuguese spoken in Portugal and Brazil. I now have slightly more sympathy for those who proposed to separate the BR and PT Wikipedias, while still strongly disagreeing with it.

Some examples from different word categories.

PT, BR (English in parenthesis)

Epidemiology: BR is closer to English.

Cancro = Câncer (EN: cancer)
SIDA = AIDS (EN: AIDS)
Varicela = Catapora (EN: chickenpox)

Computers: BR has tendency to borrow from English, while PT has a greater tendency to reinvent or rederive them.

Rato = Mouse (EN: mouse)
Ecrã = Tela (EN: screen)
Ficheiro = Arquivo (EN: file)

(tangentially, a horrible term for HD is "Winchester"; thankfully seems to be disappearing)

Animals

pirilampo = vagalume (EN: firefly)

Musical instruments:, the most confusing area. PT is closer to English.

Guitarra = Violão, Viola (EN: guitar)
Guitarra Eléctrica = Guitarra (EN: electric guitar)
Viola = Viola (EN: viola)
Acordeão = Sanfona (EN: accordeon)
Sanfona = viola de roda (EN: hurdy-gurdy)
Harmónica = Gaita (EN: harmonica)
Gaita = Gaita de Foles (EN: bagpipe)

See more on Wikipedia.

[]

Thu, 26 Nov 2009 22:40:29 GMT

The plus side of doing PhD applications again is that I'll recognize a lot of new people at NIPS, and know what they work on.

does promotion associate = tenure ?

Thu, 19 Nov 2009 08:35:55 GMT

In which US+Canadian universities does promotion to "Associate Professor" imply tenure?

Make your contribution here: http://www.optimizelife.com/wiki/Does_promotion_imply_tenure

Carl Zimmer

Thu, 19 Nov 2009 07:44:36 GMT

Today I saw Carl Zimmer talk about the evolution of viruses, specifically how they cross the species barrier (between birds, pigs and humans). Although birds seem to be the (direct or indirect) source of all human viruses, and even though they harbor all the Hs from 1 to 16, it is pigs that seem to be good at promoting "viral sex", i.e. recombinations between different virus types. Apparently, industrial-scale pig-farming barns seem to be huge labs for viral creativity.

There is concern that the 60%-deadly H5N1 might evolve the ability to spread from human to human, but it still seems to be 13 mutations away from doing that. I found this number awfully precise, and upon questioning, he told me that these 13 genes are the ones that are always present when you look at the 'diff' between bird-virus and human-virus, i.e. right before and right after it migrates to humans. The "always", however, only refers to the great pandemics, of which there are 3... so I'm concerned about the statistical significance of this.

He showed us this awesome video from NPR (who knew they made videos!):

QOTD: "The virus lives in the tropics year-round, and comes up North during our winter. If it's any consolation, they come here to die"

hash tables and equality definitions

Wed, 18 Nov 2009 22:36:59 GMT

I've just realized that using hash tables properly requires a notion of when two objects have equal value.

How come I never had to worry about this in the past?

---

I want my hash table to encode an equivalence relation (permutations of node labels), automatically dealing with my label-switching problem.

---

Do any of the normal languages/libraries let you pass an equality test, and adapt the hash function accordingly? I imagine you'd need to pass a function that brings the objects into a normal form.

---

I've installed R's hash library, but it seems to be pretty basic.

viruses vs the immune system

Wed, 18 Nov 2009 10:59:56 GMT

I'd love to see a plot showing "number of viruses exposed" vs "probability of infection", where subjects are injected with viruses in varying quantities.

Presumably a single virus (what do you call an individual virus?) isn't enough to make you sick most of the time. I imagine we'd be sick all the time if that were the case...

If you inject a single virus into bloodstream, is it most likely to get lost, get blocked or reproduce?

Can we develop immunity by low-grade exposure to a virus?

My question also applies to other microbial parasites.

wind

Tue, 17 Nov 2009 10:55:00 GMT

Wow, this storm is so strong that I'm fearing for my electricity. TheWeatherNetwork says 35km/h but has a link to a Wind Warning:

WIND WARNING: Greater Vancouver Issued at 8:35 PM PST MONDAY 16 NOVEMBER 2009
SOUTHERLY WINDS 50 TO 70 KM/H WILL DEVELOP LATER THIS EVENING. THIS IS A WARNING THAT POTENTIALLY DAMAGING WINDS ARE EXPECTED OR OCCURRING IN THESE REGIONS.

Is there a cheap and easy way to harness wind power?

I'm not expecting the trees to have many leaves left in the morning!

Gibbs sampling

Fri, 13 Nov 2009 12:26:27 GMT

According to the sources I've read, Gibbs sampling doesn't really tell you how to sample... it's just divide-and-conquer: you still need to sample each variable from its full conditional.

In my problem, that would be more trouble than it's worth, since no denominators are canceling. So I'm doing good old Metropolis-Hastings, with a proposal that makes its trajectory resemble Gibbs: it randomly chooses one variable to modify, and then randomly chooses a value for that variable. In other words, the proposal is uniform over the neighbors in the "generalized hypercube"^*.

I can easily compute the posterior. In traditional MCMC, I think you would weight the samples by how often they appear. But doesn't it make more sense to directly compute the posterior in all sampled models?

Should I be using MCMC at all?
How else am I going to find the set of high-probability models? (Maybe what I want is a mode-oriented stochastic search, as my EA project did.

Believe it or not, it's my first time actually implementing MCMC.

Also, I'd like to know what proportion of the posterior mass my sampled models account for... but this is probably VeryHard to do if I can't enumerate the models (otherwise we could know whether the chain has mixed).

^{* - what do you call a non-binary hypercube? I mean: let each node be a string in {0,...,k-1}^n, and neighborhood relation is defined by set of strings that differ in exactly 1 character. When k=2, we get the n-dim hypercube. What's the word when n>2?}.

conduction, convection and infrared

Fri, 13 Nov 2009 08:40:02 GMT

Having a TV this week is putting me more in touch with ordinary folks, for whose sake science maxims are subverted, e.g. "The NuWave Oven uses all 3 forms of heat transmission: conduction, convection and infrared!!". Everyone knows that "radiation" is bad for you...

As you may know, MRI used to be called NMRI, but medicine and "nuclear" don't go together.

Facebook to Google Calendar

Thu, 12 Nov 2009 01:46:43 GMT

Has anyone succeded in automatically syncing Facebook events into Google Calendar, as soon as I select "Attend" or "Maybe Attend"?

slime mould intelligence

Thu, 05 Nov 2009 01:04:55 GMT

Does slime mould perform MCMC?
http://en.wikipedia.org/wiki/Physarum_polycephalum#Event_anticipation
http://discovermagazine.com/2009/jan/071

LaTeX figure placement

Tue, 03 Nov 2009 23:55:01 GMT

Dear LJ Genie,

Why is LaTeX so willful when it comes to figure placement? Why doesn't it listen to my "[h!]"?

Garden of Earthly Delights

Tue, 03 Nov 2009 10:26:54 GMT

After my place is painted, I will try to look for a place to a print a big poster of the 500-year-old painting, the Biblically-inspired Garden of Earthly Delights.

If it had been painted by Nostradamus, I would say he was envisioning Burning Man.