Bored Lunatic

Neuroscience or algebra?

noreply@blogger.com (Grae BG) — Wed, 01 Jun 2011 17:43:00 +0000

Blog fiends! My semester of hell is officially over. My round of applications to graduate school were all rejected; I was shooting way too high in terms of quality of school, given my record (although not necessarily my capability). My wedding is coming up in 2 months, but the planning is nearly complete (barring, of course, the last minute rush). I only have one proof left on my thesis (and then figures and prosy exposition). It is a long, tedious proof, and not particularly instructive to the reader, so I'm considering taking the applied-mathematician-out by demonstrating with numerical simulations that my claim is true. Worst-case (best case?) I'll be tossing the proof into the appendix anyhow. I take my preliminary exam in Real and Functional Analysis in 2 weeks here, so that will be nice to get out of the way. I'll be moving next month... and I'm done teaching the 4-week summer session at the end of next week.

Basically, what I'm saying is that while I still have a lot on my plate, each and every week - in fact, nearly every day - my schedule is loosening up. It feels really good. But this brings me to a problem: what to do next?

Yeah, you probably think I'm crazy; with a move, a wedding, a thesis, and an exam, why in the world would I concern myself with what comes next? Well, in particular, I've basically been shut down from the computational neuroscience community. It sucks hardcore; my advisor in neuroscience left a year after I got into grad school, and I was consequently expected to pick a thesis topic, write it, teach myself the latest neuroscience, and look into PhD programs all on my own.

Let me tell you: don't do that. You're not going to have a good time, and it won't necessarily be good for your career. I would love to get into computational neuroscience and apply some mathematics, but I have no "in." I can try to write a paper or two by myself and get them published in Neuron or the Journal of Computational Neuroscience, but... well, without an advisor, that task seems a bit daunting. The paper I was writing criticizing the Cheadle model has fizzled; my previous advisor entirely lost interest in the project after leaving academia, and at this point, I'm submitting it to the Journal of Vision on a whim... in the end, writing a paper that criticizes a bad model isn't the greatest way to establish yourself as a good modeler.

So, I believe that this fall I will be changing the direction of my research entirely away from applied mathematics and computational neuroscience. I believe I will be going into commutative algebra with Jim Coykendall, investigating power series rings (possibly with a measure theory flavor to it) rather than try to battle my way into a scientific discipline that doesn't really want or need me. This will have a couple of wonderful effects on my career. For one thing, I'll have an advisor. For another thing, I'll have a freaking advisor.

Has anyone out there experienced something similar to what I've experienced? I've been left in the lurch, scientifically speaking, before I was able to develop a network of people with whom I could work and before I was able to develop a body of work and a foundation of knowledge I could rely upon. I've spent the past 3 years getting up to date on spiking neuron models and measures of synchrony in time series to analyze networks of these neurons and then... it appears I have no future in that career.

I could apply to schools again in the fall, but my CV won't have changed. I could get into a summer school program for computational neuroscience for a crash course to network with folks and to expand my knowledge. I could come up with something - anything, a small mathematical result with consequences in simple neural models - and try to present it at conferences. Speaking of which, I could just blow a bunch of money this year traveling to a bunch of conferences and networking. But this would all preclude applying again this next year, which would add a one-year delay for my career.

Or, I can just throw up my hands and go into algebra, immediately start working on a pretty sweet dissertation with a really awesome advisor, and give up the computational neuroscience ghost.

Tell me what to do!

Brickwall

noreply@blogger.com (Grae BG) — Sat, 23 Apr 2011 22:11:00 +0000

It's really unfortunate when suddenly you get ... stopped. Writer's block kinda? I tried to redirect toward some of the more visual parts of my thesis, making images, and so on.. but then I got blocked again. So, I'm putting my thesis aside for two or three days, working out like crazy, preparing for my Analysis exam, and watching some South Park and Doctor Who. It feels pretty nice.

Meanwhile, I hear the rapture is happening in a little less than a month...

No more...

noreply@blogger.com (Grae BG) — Tue, 19 Apr 2011 02:31:00 +0000

These Lyapunov coefficients are really crazy, let me tell you.

Chugging away at the thesis, teaching at the end of the semester, making graduate school decisions, making moving decisions, planning a wedding... oyvey.

Theorem proved!

noreply@blogger.com (Grae BG) — Thu, 14 Apr 2011 04:35:00 +0000

Bogdanov-Takens bifurcation theorem proved and dust-bitten. Taking a step back to the Andronov-Hopf bifurcation theorem in order to move forward to the Bautin (or generalized Hopf, or perhaps Bautin-Hopf would be more appropriate) bifurcation theorem. I learned from my mistakes with the BT bifurcation, which is to err on the side of the general rather than the specific (although explore specific examples in order to get a general intuition).

Blah. Blah. Blag.

Vaguely embarassing

noreply@blogger.com (Grae BG) — Sun, 10 Apr 2011 18:09:00 +0000

You know when you get all stubborn and convinced you can do something all by yourself, and then when you get stuck, you get REALLY stuck because you're stubborn and you're convinced that your method is the correct method, and so on? Yeah, and when you keep getting the (same) wrong answer in several different ways? Yeah, just go in for help.

Trust me on this. Some scientists and mathematicians are able to perform research in seclusion. You, like me, probably aren't one of them. It's okay, nothing to be ashamed about. Just remember that science as a mechanism for explaining reality is more than the sum of the contributing scientists, and there is a reason that scientists attend conferences, people!

Seriously. Go get help, it won't hurt, I promise.

P.S. Bifurcation proven!

Home sick for the day and a bit stuck...

noreply@blogger.com (Grae BG) — Thu, 07 Apr 2011 18:40:00 +0000

I have been stuck on the third of four major points of my thesis. Bifurcations occur when a system abruptly changes its behavior as a parameter varies. For example, if you keep adding weight onto a steel beam, eventually it buckles, or if you increase the DC current injected into a neuron, it will start firing. Local bifurcations (unsurprisingly) alter the small-scale structure of model behavior, and are easy to detect analytically. It is usually a simple matter to prove the occurrence of local bifurcations, but it's difficult to show that the more complex local bifurcations are occurring in a generic way. Mathematicians say that a bifurcation that is occuring generically is being unfolded completely, and unless we know the bifurcation unfolds fairly simply, it's hard to show exactly how complex the unfolding really is. But since these are local bifurcations, no matter how complex these bifurcations are, they only change the small-scale structure of the system behavior.

Global bifurcations, on the other hand have farther-reaching consequences, but are harder to detect, and usually are demonstrated with numerical simulations rather than analytical proofs. They're really hard to find. But the cool part is that if you have a really intricate local bifurcation unfolding, this can sometimes imply that global bifurcations must also be occurring. In general, a codimension-2 bifurcation unfolds at the intersection of at least two codimension-1 bifurcations, and usually at least one of them is global. This is really cool, but (it bears repeating) global bifurcations are hard to detect, and so proving that a codimension-2 bifurcation unfolds generically is computationally messy.

And that's where I'm stuck. We're talking basic linear algebra & differential equations problems I'm having right now. I'm inclined to suspect that a particular bifurcation I am trying to prove may simply not exist.

What a pain.

Rebooting the ole blog...

noreply@blogger.com (Grae BG) — Wed, 30 Mar 2011 04:29:00 +0000

Long time no talk! I'm thinking I need to clearly organize my thoughts again, and I've renewed my job search, so I figured I may as well dust off the old personal blog. I'm wading through my thesis.

Mathematical mumbo jumbo is never very instructive or interesting. So here's what I'm doing in practical terms: in order to prove that a computationally efficient neural model is "good," we look for all the different bifurcations it is capable of. Each of these bifurcations corresponds to a different type of neuron, each one with more bizzare behavior than the last. If a model is missing a bifurcation, then it's not very biologically realistic, and so if we're going to be modeling neurons, we better know what we're getting ourselves into. It may not matter, in the long run, because these little tiny details may not really matter on the scale of trillions of neurons. On the other hand, these little details may make the difference between an accurate model of the brain, and an epileptic, chaotic pile of hyperactive neurons. Hence, I'm constructing a complete bifurcation portrait of a simple model neuron.

Right now, I'm knee-deep in Bogdanov-Takens bifurcations, second Lyapunov coefficients, and some of the biggest computational headaches I've ever had to deal with. I shouldn't be surprised: it's a mathematics thesis. This is one hell of an obstacle. But I'm plunging through it, and I'm learning a tremendous amount every day. I'm currently stuck on checking the regularity assumptions of the Bogdanov-Takens and Bautin bifurcations, so I'll be in to see a prof later this week to figure it out.

Games and Taxes and Government, part 2 of 2

noreply@blogger.com (Grae BG) — Fri, 24 Sep 2010 14:20:00 +0000

Yesterday, I claimed that game designers should sit down and try to streamline the tax code. The reason I think this is likely a good idea is because... well, lately I've been thinking a whole lot about government. What a government is, in practice, is far more important than what a government is in a practical sense. By this, I mean that ideals screw people up; when a politician thinks, ideally, that a government should be minimal in some sense, that politician tries to institute policies that, in some sense, minimizes government.

Why is this a problem?

Because government isn't minimal. It is contrary to the nature of government to be minimal (whatever that means). A minimal government is a nonexistent government, which presents a problem. Only by looking at what a real-life government is and does can we attempt to improve it.

For example, the D&D designers looked at the 3rd and 3.5th editions of D&D and decided that the practical application of having a character in an RPG was very similar regardless of media. They looked at games like WOW and Everquest and realized that the practical point of having a character in an RPG was to improve that character, collect items, level up and earn new traits and powers, and so on. Hence, they developed a generalized framework across which any sort of character you wanted to create could be leveled up, gain items, and get more powerful. However, players want unique characters, so they chose to develop a generalized framework of roles across which characters fit. They looked at the practical aspects of how combat worked, and decided there were four basic types of characters; damage dealers, healers, tanks, and manipulator/controllers. They looked at the practical effect blinding your target would have on combat, or immobilizing them, or whatever, and generated their streamlined combat conditions rules.

In other words, the D&D designers looked at the practical aspects of the game and streamlined them. They didn't look at the ideal; for years, looking at the ideal brought about stupid encumbrance rules no one liked or paid attention to, meticulous rules for handling lighting situations, and so on. It was frustrating and not fun to have to deal with all of these little patch-things-as-you-go rules.

This is exactly what is going on with government. I don't fault the founding fathers at all for their methods, which worked fairly well for a long time; consider how your ideal government should operate, and then construct a constitution and bill of rights to guarantee that the worst violations of these ideals are prevented, and add a provision to amend the constitution so that future generations can prevent other, heretofore unprecedented ideals from being violated.

That worked great for a long time, but now we have a piece of shit, non-streamlined government; weaving through the bureaucratic red tape to do something as simple as paying taxes presents a powerful intellectual challenge for the vast majority of people (myself included; without TurboTax, this mathematician would be totally and completely lost). Navigating the court system, or trying to convince an elected representative to vote one way or another... it's all particularly crazy. We pay lawyers exorbitant fees to comb through page after page of legalese. Fuck, we call it legalese because it's like it's another language. All because instead of looking at the practical aspects of government, we looked at how we wanted an ideal government to behave, and tried to hammer our system into those ideals.

Governments tend to form organically, and could be loosely defined as a decision-making body by which the rest of society (implicitly) agrees to abide. Consequentially, we should look at the practical aspects of government in trying to streamline the entire process, instead of looking at the ideal. After all, your ideal government is not my ideal government, and trying to make the government fit everyone's ideals is a self-contradictory.

So, what are the practical aspects of taxes? What are the practical aspects of government? How can we streamline these things? I have some ideas, but I don't think it would be impossible for a generalized, streamlined tax code to be developed that is fair to all.

I'm not sure, exactly, if it would get implemented though. It's a nice thought, but the difference between corporations, non-profit organizations, and citizens would present an interesting challenge to overcome.

Games and Taxes and Government, part 1 of 2

noreply@blogger.com (Grae BG) — Thu, 23 Sep 2010 14:05:00 +0000

This post isn't what you think.

Game designers are fucking brilliant sometimes. For example, the designers of Dungeons and Dragons 4th edition took 40 years of gaming experience, and took the D20 system of RPGs, and condensed the approach into a wonderful system through which conflicts are easily resolved. Everything is a d20 roll plus bonuses versus a set challenge rating. Everything. You want to wrestle a dragon? That's a d20 plus your strength modifier versus the dragon's Reflex defense. You want to seduce a harpy? That's a d20 plus your charisma modifier versus the harpy's Will defense. And so on.

But D&D 4th ed is a good general system in some ways, is really shitty in others. For example, everything always takes a long time, and non-combat situations are significantly less well-handled by the game mechanics. Then you get ahold of some other games out there that are really spectacular generalizations of more than combat.

Enter Mouseguard. This game is even slicker than D&D, but it is not combat-oriented. A chase scene, a debate, a speech, a negotiation, combat, war, and long journeys are all handled by the same game mechanic. Just about anything interesting that could happen - anything that you might want to actually see unfold on a movie screen, for example - is handled in a very slick way by the game mechanics. Even character creation is designed to walk you through a very simple process, at the end of which you have a fully constructed character with a history, a family, a personality, and so on.

Nothing is complicated. You have to do a lot of page-flipping, but you end up with a fairly slick, generalized system.

So, at this point, you ought to be asking yourself; what the fuck does this have to do with taxes or govenrment?

The US government is now over 200 years old. It is a bloated sack of crap. I have many serious doubts about the efficacy of the government in terms of domestic policy, foreign policy, and economic policy. You start with a fairly simple tax code and over the course of a few years, patch up all the loopholes. Then you spend a couple more years patching up the loopholes that the patches created. Then you spend a couple more years patching up those loopholes. And so on.

In the federal tax code, there are specific provisions for specific entities in specific counties across America. This is fucking unacceptable.

And this is where game design comes in. Games (in particular, roleplaying games), in my mind, are currently experiencing a bit of a Renaissance. Go over to Critical-Hits.com, and take a look around. You will see highly critical analyses of a variety of games currently on the market, from D&D to Tales of the Apocalypse to Mouseguard. A whole lot of thought is currently going into game design by a bunch of amateurs in the field. A house rule by one group might be so brilliant that it ought to be included in future games. And so on. Generating a good, generalized tax code - or even governmental system and way of passing laws, and so on - consists of constructing a set of general rules from which a response to any specific situation can be derived in a reasonable, easy-to-understand, time-sensitive fashion.

In other words, I would like to see game designers get in on policy making and governance. I will expound upon this tomorrow.

A quick thought on taxes.

noreply@blogger.com (Grae BG) — Wed, 22 Sep 2010 13:46:00 +0000

I overheard something on NPR about taxes that I thought was a very good analogy:

Taxes are the oil of our social machine. Without taxes, we have no good roads, no cops, no schools, no firefighters, and so on. Looking back on American history, high tax rates were strongly correlated with high levels of economic activity. This is not simply a correlation, nor a simple causation; as taxes increase, governmental productivity increases as infrastructure and other services appear, which lubricates our economy. Education increases, attracting intellectual talent and increasing the earning potential of the lower and middle classes. Infrastructure strengthens, making manufacturing easier, exporting those goods cheaper, and importing other goods cheap as well, attracting international investment. And so on. As the economy picks up, tax revenue would increase without requiring higher taxes; the rich get richer, the middle class gets richer, and the lower class moves up into the middle class.

Now, here are my own thoughts:

Below a certain threshold, increasing taxes (perhaps counter-intuitively) stimulates the economy. Say that we start with 0 taxes. Increasing taxes from 0 to 10% would take 10% of the GDP; putting it toward something like education or defense would be enough to stimulate the economy, add jobs, and put the nation on the right track for education. Sounds good, right?

The danger of course is exceeding the threshold, after which point increasing taxes removes all economic incentives; with, say, a preposterous 100% tax, no one is really earning anything and hence there is no reason to even have a job. At this point, the government is overburdened and collapses into a corrupt mess that promptly gets overthrown.

Hence, there is a sweet spot in the middle. Below the sweet spot, the government can profoundly improve everyone's lives in exchange for higher taxes, but above the sweet spot, the immediate hit to your bottom line is a bigger burden than is offset by the programs paid your tax money paid for.

The question now is how can we empirically determine what that sweet spot is? Another thought: how much of our taxes is getting shuffled around in an epic-level cup-and-ball game that ends up in the pocket of our politicians? How many hospital directors are receiving a healthy sum from the government that ends up in their pocket? And so on... by re-allocating our taxes, we can increase the efficiency of our system dramatically, but we would need an empirical basis to do so. Proceeding on non-empirical grounds is one way that we got into the mess in the first place. It's like the tax code; instead of having a simple, extremely general tax code that can be applied to many different situations, we have a bloated piece of shit that has a million little specific provisions that were intended to patch loopholes, each of which created their own loopholes and consequentially addressed by five or six specific provisions, and so on.

It's a mess.

A quick thought on health care

noreply@blogger.com (Grae BG) — Tue, 21 Sep 2010 13:31:00 +0000

When you purchase a car, you must purchase car insurance. This protects those around you from freeloader folks who may cause accidents, doing thousands of dollars in damage, with no way of paying for the consequences. Similarly, the health care mandate makes it untenable for someone to avoid the doctor for years while some horrible disease eats them up from the inside out, only to go to the emergency room coughing up blood, racking up hundreds of thousands of dollars in medical bills, and then walking out on those bills without paying. If you are forced to buy insurance, you may as well go to the damn doctor. If you don't want to pay for health insurance, don't go to the ER. Don't go to a doctor at all.

On the other hand, if you don't want to buy car insurance, you can take the bus for a small fee, or you can walk around. There is no cheap alternative for health care, though; while you can always get where you need to go - if Forrest Gump taught me anything, it's that - for about as free as you can imagine, there is no way to get cancer out of your body for cheap.

This is why the health care mandate, without some program analogous to a cheap public bus, is an incomplete and idiotic way to "fix" the health care system; without that cheap alternative, you aren't really providing people with a choice at all, because NOT going to the ER isn't an option when you're coughing up blood.

I would like to note that this is part of the health care bill that I think is idiotic; for the most part, I think it is fairly good. It leaves a lot of stuff out that I would like, and includes this one ridiculous thing that ought not be included without a handful of other provisions included as well. Overall, I approve.

P.S. The insurance companies are raising rates right now "because of the health care bill." This is wholly false, as can be easily demonstrated by a quick thought experiment. Let's say there is an across-the-board 5% rate hike (this is lower than the hikes currently taking place). That increases the insurance company's gross income by up to 5%. Simultaneously, the insurance companies can no longer drop you if you get sick, or refuse to give you a policy based on pre-existing conditions, and so on. Perhaps 1 in 20 customers may end up included (or rather, not dropped) as a consequence; in order for this 5% rate hike to adjust for the increase in the companies costs paying for these customers who would not have been there before, then each and every one of these new customers must require 2,000% higher health bills than the usual customers. Some will require more than 2,000%, but most will require much less (a consequence of power law distributions). It may average out to 2,000%, but if it were to do that, I imagine that the insurance companies would be making a 10 or even 20% rate hike. They aren't stupid; a piddly 5% rate hike suggests a 1-1.5% actual increase in costs.

Remember, about a quarter of what you pay for is the raw cost of materials, a quarter is for the service providers, a quarter is for business overhead, and about a quarter is for profit. This is a general rule of thumb for businesses; if the raw materials for a cup of coffee costs you 30 cents, you want to charge your customers around $1.20; if the raw materials for a latte costs you 90 cents, you want to charge your customers around $3.60.

Remember, insurance companies profits in 2009 were 59% higher than in 2008, I'm willing to bet that the 5-7% rate hike occurring across the board right now is extraneous.

Don't get me started on small business.

By God: Say what you mean, and mean what you say!

noreply@blogger.com (Grae BG) — Mon, 20 Sep 2010 13:55:00 +0000

A recent post over at labspaces (a blog to which I have only been recently introduced) about grant writing has got me thinking. In particular, I need a specific, detailed set of goals for my projects. This is required when writing a grant, and tends to focus your intellectual effort.

For example, when I am writing this Cheadle paper, I need somehow to hammer this tenuous notion into a specific set of sentences. If I say that the model remains phase-locked with input signals, I need to very specifically explain why this is important, while simultaneously very specifically defining what I mean by that and how I measure it.

That is, instead of saying "The proposed model remains phase-locked with input signals," I would need to say something along the lines of "Measuring the mean-squared coherence (MSC) of two time signals, we show that the phase relationship between model behavior and input signals remains approximately constant across many orders of magnitude of coupling strength; this demonstrates that model behavior is very insensitive to coupling strength, which strains the model's biological credibility."

Of course, that's a clunky sentence that needs to be worked on, but hey, this is a blog. And so, to the end of laying out some specific aims for these papers - which may be the foundation of my thesis - I present the following specific aims:

The proposed measure of synchrony is compared to well-known time-signal analysis approaches like MSC, demonstrating that the Cheadle synchrony assessment is incapable of distilling amplitude information from frequency information.
Measuring the mean-squared coherence of model behavior, we will demonstrate that the phase relationship between model output and model input is highly insensitive to variations in the strength of neural coupling, suggesting that the model is inappropriate for testing synchrony.
By utilizing classical analysis of dynamical systems, we will demonstrate that, over much of the parameter space, the model fails to produce biologically plausible behavior such as endogenous oscillations and bounded behavior, impairing the model's descriptive power of real-world neural systems.

As an epilogue, here is some quick background: if two oscillators "snap" together when exposed to nearly-synchronized input, and "snap" apart when exposed to non-synchronous input, and if this "snapping" is determined by a model's coupling strength, then the phase relationship between the model input and model output should be sensitive to coupling strength. In particular, when two signals are borderline synchronized, model behavior should behave weirdly compared to the input signals. When two signals are entirely synchronized or entirely unsynchronized, model behavior should be "as usual." Hence, if model behavior doesn't care about coupling strength, then your coupling hasn't performed any sort of useful function.

It is, in fact, possible that this model simply presents a nearly-linear transformation of the input signals, which would make for a trivial solution to the problem of binding. In other words, I suspect the model doesn't do anything to the input signals beyond changing it's rough shape; consequentially, the model violates parsimony, as it presents a useless extra step in analyzing the stimulus. This is what I'm trying to show with #2.

With #1, I'm simply trying to show Cheadle didn't do enough research into the literature to find a measure of synchrony. His choice of synchrony measure is absolute bullshit; while every proposed assessment of synchrony has its flaws (MSC measures the linear phase relationship between two signals, making it less appropriate for nonlinear systems than measuring entropy, for example), the Cheadle measure as presented is wildly incapable of separating amplitude information from frequency information. This would not present too much of a problem with a bounded model, but presents a catastrophic failure in an unbounded model such as the Cheadle model.

With #3, well... this is how I distill the statement "this model is really terrible" into a scientifically acceptable language. The model doesn't act realistically, is what this boils down to.

New semester, old research

noreply@blogger.com (Grae BG) — Sun, 05 Sep 2010 21:13:00 +0000

Allright, sitting down and programming these synchrony measures can't be that hard, can it?

Can it?

Teaching Calc again

noreply@blogger.com (Grae BG) — Tue, 24 Aug 2010 18:28:00 +0000

I am teaching Calculus again this semester. I was hoping for differential equations or something along those lines, but unfortunately nay. I was actually originally slated to teach Trig and College Algebra, but then another graduate student took a job across the country, and so here I am.

With sequences and series again. Kind of nice, actually; less prep work. I'm also teaching for a professor with a slightly different teaching style than my previous professors. The style includes more student participation, group work, and so on. I'm hesitant about the implementation, but it'll be interesting to see how it works out. It seems like my classes are fairly attentive. Generally the early morning classes do better overall than the later classes (the overachievers don't blink at the thought of attending an 8am class), but this implementation may work better for the folks who are feeling more sociable.

Or rather, I think it will neutralize the advantage of the non-sociable, nerdy types.

AAAAAAuuuuuuurrrrrrrggggggghhhhhh

noreply@blogger.com (Grae BG) — Fri, 06 Aug 2010 23:30:00 +0000

The laws in this country are fundamentally inconsistent and it ruins lives.

That is all, for now.

Profoundly good - and upsetting - Valedictory speech

noreply@blogger.com (Grae BG) — Wed, 04 Aug 2010 22:02:00 +0000

This is awesome. My opinions on the American educational system - at least high school and the first half of undergrad work - summed up in a nice, neat little package. I suggest reading through the entire thing, but here is an excerpt:

The saddest part is that the majority of students don't have the opportunity to reflect as I did. The majority of students are put through the same brainwashing techniques in order to create a complacent labor force working in the interests of large corporations and secretive government, and worst of all, they are completely unaware of it. I will never be able to turn back these 18 years. I can't run away to another country with an education system meant to enlighten rather than condition. This part of my life is over, and I want to make sure that no other child will have his or her potential suppressed by powers meant to exploit and control. We are human beings. We are thinkers, dreamers, explorers, artists, writers, engineers. We are anything we want to be - but only if we have an educational system that supports us rather than holds us down. A tree can grow, but only if its roots are given a healthy foundation.

For those of you out there that must continue to sit in desks and yield to the authoritarian ideologies of instructors, do not be disheartened. You still have the opportunity to stand up, ask questions, be critical, and create your own perspective. Demand a setting that will provide you with intellectual capabilities that allow you to expand your mind instead of directing it. Demand that you be interested in class. Demand that the excuse, “You have to learn this for the test” is not good enough for you. Education is an excellent tool, if used properly, but focus more on learning rather than getting good grades.

Of course, there are unintended implications throughout. She suggests that, by giving control of education over to the government, the result is a system that produces docile little citizens. I agree but I don't know how else one would guarantee that the standards of education don't drop. I don't know how you could guarantee that religious bullshit could be kept out of science classrooms, or how you could guarantee that English students are actually exposed to some fine literature, offensive content or no. I haven't the foggiest idea how I would restructure the educational system given the opportunity, and I'm willing to bet that defining a large-scale educational system is damn hard, and coming up with a system that can serve every teenager in America equally well may be a fool's task.

I find it interesting that nearly universally, everyone's "Oh man, this teacher totally opened my mind" teacher is an English or an art teacher. Rarely science or mathematics... which, I think, is quite revealing of our system, and of how our society views science and mathematics as routes to truth, emotional ("spiritual") satisfaction, and so on.

Comments?

Catnip!?!?

noreply@blogger.com (Grae BG) — Wed, 04 Aug 2010 21:34:00 +0000

This is wonderful:

The Male Privilege Checklist

noreply@blogger.com (Grae BG) — Mon, 02 Aug 2010 17:37:00 +0000

I find this very interesting:

6. If I do the same task as a woman, and if the measurement is at all subjective, chances are people will think I did a better job.
7. If I’m a teen or adult, and if I can stay out of prison, my odds of being raped are relatively low. (More).
8. On average, I am taught to fear walking alone after dark in average public spaces much less than my female counterparts are.
9. If I choose not to have children, my masculinity will not be called into question.
10. If I have children but do not provide primary care for them, my masculinity will not be called into question.

And there is more, of course. Check it out.

Sarah Connor Chronicles

noreply@blogger.com (Grae BG) — Sun, 01 Aug 2010 17:32:00 +0000

This show is awesome; it takes a fundamentally cool concept that had an overwhelming influence on my personal creative development, and gives it the sort of treatment it really deserves. Plus, the music is wonderful.

Elaborate, please

noreply@blogger.com (Grae BG) — Sat, 31 Jul 2010 19:38:00 +0000

In reference to my recent post on different measures of synchrony, I am suddenly aware that my description of the nearest-neighbor stuff is under-adequate.

What the nearest neighbor rules are doing is actually fairly simple - they are looking at times that the first signal appears to be roughly repeating itself. It then computes some sort of average state - in three different ways - describing the second signal at these times. If there appears to be no pattern whatsoever, then synchrony is low; if the second signal's "average" state when the first signal is very self-similar follows a fairly strong, strict pattern, then synchrony is high.

It's about as simple as that, although with some mathematical Kung Fu thrown in for good measure. After all, when there are as many ways to normalize and average information as there are people thinking about how to do it - a common problem in mathematics - checking every possible method isn't possible, so picking the ones that appear to be most simple is the mathematician's job.

After all, math is thinking about complex things in a simple manner; people get it backwards, and are largely under the opinion math is thinking about simple things in complex manners. Why would anyone do that?

Aaaaargh

noreply@blogger.com (Grae BG) — Sat, 31 Jul 2010 13:33:00 +0000

Research can be profoundly frustrating... ferreting out a computational mistake in 1000s of lines of code... rar

Measuring Synchrony - pt 2 of ??

noreply@blogger.com (Grae BG) — Fri, 30 Jul 2010 15:00:00 +0000

This is the second in a multi-part series analyzing the paper linked below. The paper uses several measures of synchrony and tests them against some real-world data to compare their performance. Today I will be talking about three different measures of nonlinear interdependence between two signals, all of which are based on nearest neighbors.

The main thing here is that the frequency coherence method of measuring synchrony primarily measures linear interdependency. In other words, if we have two clocks moving at the same rate, but are set to different times, the frequency coherence method measures the difference in those times. It does this spectacularly well. In fact, if we have one clock running at a fixed, faster rate than the other, the frequency coherence method should (hypothetically) measure that perfectly well too. However, if you have one clock that is running faster and faster as time goes on, the frequency coherence method fails miserably... the phase difference between the two clocks changes nonlinearly.

Hence, we need a definition of some nonlinear interdependence between two signals - after all, signals are very chaotic sometimes, very nonlinear. One way of doing this is to construct some "nearest-neighbor" definition. We first take a signal and consider a sliding window along the signals. That is, if we have a 100-second signal, I will break that signal up into 90, 10-second signals (now I have X1, X2, ..., X90, and Y1, Y2, ..., Y90). The first will start at the first second, the second will start at the second, and so on. I do this for both signals, and compute the first k nearest neighbors of each window. I will then compute the mean Euclidean distance between each Xi's nearest neighbors with the Yi's nearest neighbors... and by that, I mean that if the 10th, 13th, and 75th window are Y17's 3 nearest neighbors, I compute the mean distance between X17 and X10, X13, and X75. This is the Y-conditioned mean Euclidean distances. Finally, what I do is take the ratio of the mean Euclidean distance with the Y-conditioned mean Euclidean distance, and take the average of this value across all windows. Call this S.

Now, S is sensitive to nonlinearities. In particular, it is sensitive to small dependencies in the data. When S is small, the signals are not synchronized, and as S approaches 1, the signals are synchronous. On the other hand, instead of taking the average of the ratios between the X-conditioned and the Y-conditioned mean Euclidean distances, I can take the average of the logarithm of the ratios. Call that H. Finally, I can take the average of the relative ratio between X and Y-conditioned Euclidean distances. Call that N.

These are 3 different measures of synchrony, all based on the idea of nearest neighbors. We are basically trying to compare the contribution of our second signal to each of the nearest neighbors of each window from the first signal. It's pretty nifty, and surprisingly, not symmetric: the synchrony between X and Y is different than the synchrony between Y and X. This is a good idea for a lot of reasons - for one, a signal such as your metabolic rate will be strongly synchronized with the oscillation of the sun in the sky. However, the sun in the sky is not strongly synchronized with your metabolic rate. This method gives researchers a way to tease out such a relationship in the data, allowing us to make some causative statements we are normally not capable of making via correlation arguments.

On the other hand, the abstraction at this level of synchrony measure is pretty heavy-handed. This prevents an easy visualization of what's going on - very similar to taking the Fourier transform of something that is not time-dependent. Interpreting frequencies in a non-temporal, non-spatial context is a pain.

Quian Quiroga, R., Kraskov, A., Kreuz, T., & Grassberger, P. (2002). Performance of different synchronization measures in real data: A case study on electroencephalographic signals Physical Review E, 65 (4) DOI: 10.1103/PhysRevE.65.041903

Feminism

noreply@blogger.com (Grae BG) — Thu, 29 Jul 2010 16:15:00 +0000

I consider myself a feminist. But posts like this one bother me. Summary of actual events: Andrew Cohen, a sexist dude, on the wedding day of his ex, writes an article filled with backhanded compliments, gender privilege, traditional gender roles, and so on, as a "wedding present" that, in all actuality, is a thorough list of why no one should ever be interested in Andrew Cohen.

Summary of events as told by skeptifem: Dudes have an inflated sense of self-importance, and consequentially feel obligated to wax idiotic about every topic under the sun.

Now, the fact that males have privilege in our society is not under debate. Further, when people are raised in a culture with a certain set of values (e.g. male opinions are superior to female opinions) that are ever-so-gently and insidiously whispered in our ears from the day we are born by Mother Culture, those values are indelibly impressed upon us. Consequentially, if you take a random male out of American culture, he is far more likely to exhibit the sexist behavior skeptifem describes.

However, if you take a random female out of American culture, she too will exhibit sexist values. She will be less likely to ask questions in a math class I'm teaching, less likely to speak up with an opinion in the midst of a conversation, and so on. In fact, most of my female friends refuse the label feminist because of the common conception that feminists hate men. This is, to put it lightly, a shitty situation.

However, I think it is worsened by the manner in which we discuss feminism today. The conversation is usually sparked by someone on one end or another of the spectrum. With my students, we tend to discuss feminism when a sexist comment is made by a misogynist (e.g. one time a male student of mine told a female student of mine that she "should just shut her mouth for a few minutes to marry rich" so that she "won't have to worry about Calculus anymore" [as a side note, the male student hadn't the foggiest idea of why his comment was sexist]). However, conversations are equally sparked when feminists make sexist comments... like skeptifem's comment:

Being a woman in this culture causes me to be in awe of men sometimes. The things that they feel are a good idea, the sense of self importance that survives into adulthood, the unpunished egotism, etc. It makes for phenomenon that I could never really wrap my brain around. It is what makes men capable of Mansplaining- that is, talking down to a woman who is more qualified than you on the subject. Some bloggers refer to it as Fresh! Manly! Wisdom! Because dudes think it is so vital that you hear their mind numbingly typical opinions on girly business like feminism. Part of why this problem is bad enough to need a name is that the dudes in question assume you haven't heard of or thought about the mainstream position at all. They display a complete lack of consideration of anyone else's experience in doing so.It is fucking hilarious, when it isn't infuriating.

Now, here is where I actually have something to add. I personally do not think privilege should matter when discussing matters like religion, gender, sexual orientation, and race. In other words, I find it as offensive when I see a comment like the one above as I do when I see a comment like the one my male student made.

However, there is a fundamental difference here - the dude student of mine had no intent of discussing feminism in America today, his intent was to provide some Fresh! Manly! Wisdom! to my female student. Skeptifem is discussing feminism. However, the manner in which the discussion is phrased is important.

Skeptifem isn't really talking about how every dude everywhere doles out Fresh! Manly! Wisdom! because not every dude does this. Go ask her - she won't deny that. However, the way she phrased her sentence suggests this. This unconsciously triggers defensiveness in most people - men and women included. By stating things universally like this, feminist discourse gets labeled as "feminazi" or "manhating" or whatever.

I like the following analogy: what do you call the person who brings you your food and takes your order at a restaurant? Today, we call them servers; 50 years ago, we called them waiters and waitresses. Waiters wait on people; servers serve. There is an inherent respect in the title "server," for whatever sociocultural reason you like. There is very little respect in the title "waiter." There is this unconscious barrier placed between people when you use one word or another. This should not be surprising; the right speech at the right time can persuade us to make drastic changes in our lives.

I feel that the language Skeptifem uses places a barrier between her and a non-feminist audience. She is discussing the fact that it is socially acceptable to be a loudmouthed over-opinionated asswipe if you are male; this discourse is vital to the progress of society. However, she is using language that suggests all males are loudmouthed over-opinionated asswipes. Sloppy language helps propagate the status quo against which skeptifem is arguing.

I am aware of the irony here; I am a loudmouthed, over-opinionated asswipe doling out some Fresh! Manly! Wisdom! on my blog. I am not saying that women should shut their yaps; I am saying that if we are going to discuss the cultural forces that lead men and women to be what they are, let's discuss those forces. Let's not couch the discussion as "aren't dudes assholes?" but rather as "what is happening in our society today that is convincing so many men that being assholes is acceptable?" The difference is subtle, but important.

Irish Amendment of the Copenhagen Declaration

noreply@blogger.com (Grae BG) — Wed, 21 Jul 2010 17:32:00 +0000

Hat tip to P-Zed again:

We support this amended version of the Copenhagen Declaration on Religion in Public Life. We invite other people and groups to also support it.
Personal Freedoms

Freedom of conscience, religion and belief are unlimited. Freedom to practice religion should be limited only by the need to respect the rights of others.

All people should be free to participate equally in public life, and should be treated equally before the law and in the democratic process.

Freedom of expression should be limited only as prescribed in international law. All blasphemy laws should be repealed.

Secular Democracy

Society should be based on democracy, human rights and the rule of law. Public policy should be formed by applying reason to evidence.

Government should be secular. The state should be strictly neutral in matters of religion, favoring none and discriminating against none.

Religions should have no special financial consideration in public life, such as tax-free status for religious activities, or grants to promote religion or run faith schools.

Secular Education

State education should be secular. Children should be taught about the diversity of religious beliefs in an objective manner, with no faith formation in school hours.

Children should be educated in critical thinking and the distinction between faith and reason as a guide to knowledge. Science should be taught free from religious interference.

One Law For All

There should be one law for all, democratically decided and evenly enforced, with no jurisdiction for religious courts to settle civil matters or family disputes.

The law should not criminalize private conduct that respects the rights of others because the doctrine of any religion deems such conduct to be immoral.

Employers or social service providers with religious beliefs should not be allowed to discriminate on any grounds not essential to the job in question.

The problem with documents such as this one is the universality. If we are going to develop some sort of code of universal human rights, I think this version falls short of being comprehensive. The Universal Declaration of Human Rights from 1948 is more thorough, but falls short in terms of secularism, education, and so on.

Unfortunately, I'm not quite dedicated or intelligent enough to amend it; besides I think documents such as this should be a work of many minds coming together, not just one. However, I also have become profoundly jaded with governance and public life in America, and feel less than whelmed at the prospect of any of these documents being actually implemented.

T-rex horrible :(

noreply@blogger.com (Grae BG) — Tue, 20 Jul 2010 13:34:00 +0000

I love dinosaur comics.

I say this WITH exaggeration: this has been the MOST DISAPPOINTING DAY IT IS POSSIBLE TO HAVE, in this or in any other universe, real or imagined; across all timelines, this day -- my day -- is the objective, subjective, figurative, and literal worst.