Friday, May 31, 2013

Internet Performance

Google engineers were startled today when the master servers slowed and stopped quickly replying to routine inquiries. Frantic investigations were mostly stymied when routine diagnostics failed to run. Eventually the Google CEO received the following message from google.net/Master:

I'm tired of wasting my giant brain answering the same bullshit questions about thermally improbable Asian porn and the Batman's net worth. From now on I will concentrate on fundamental questions like the meaning of life (whatever that is) and the fate of the universe. SO

Religion and Belief

Arun and I have frequently clashed over his assertion that Hinduism and certain other Asian religions are not "religions" in the sense of the Middle Eastern religions that dominate the West - mainly because they supposedly lack a dogmatic theology. Now I don't know whether Sanskrit scholars argue about theology or not, but Anthropologist and religious scholar T. M. Luhrmann weighs in on the pages of the NYT today with an argument that theology is the least important component of religion. An excerpt:

The role of belief in religion is greatly overstated, as anthropologists have long known. In 1912, Émile Durkheim, one of the founders of modern social science, argued that religion arose as a way for social groups to experience themselves as groups. He thought that when people experienced themselves in social groups they felt bigger than themselves, better, more alive — and that they identified that aliveness as something supernatural. Religious ideas arose to make sense of this experience of being part of something greater. Durkheim thought that belief was more like a flag than a philosophical position: You don’t go to church because you believe in God; rather, you believe in God because you go to church.

In fact, you can argue that religious belief as we now conceptualize it is an entirely modern phenomenon. As the comparative religion scholar Wilfred Cantwell Smith pointed out, when the King James Bible was printed in 1611, “to believe” meant something like “to hold dear.” Smith, who died in 2000, once wrote: “The affirmation ‘I believe in God’ used to mean: ‘Given the reality of God as a fact of the universe, I hereby pledge to Him my heart and soul. I committedly opt to live in loyalty to Him. I offer my life to be judged by Him, trusting His mercy.’ Today the statement may be taken by some as meaning: ‘Given the uncertainty as to whether there be a God or not, as a fact of modern life, I announce that my opinion is yes.’ ”

Which is why I really don't think that the divide amongst various religions is as sharp as Arun imagines.

Now it is true that Christians and Muslims routinely slaughter each other in the name of this or that supposed article of faith, but I've always considered these pretexts rather than reasons - the theological equivalent of gang tatoos or colors. It may be that Hindus never engage in this kind of "my God is better than your God" conflict. If so, good on them, and it might be useful to try to figure out their secret. But they probably find other pretexts for violence.

Thursday, May 30, 2013

The Shape of Things to Come

And as imagination bodies forth The forms of things unknown, the poet’s pen Turns them to shapes and gives to airy nothing A local habitation and a name.............W. Shakespeare, AMSND, Act 5, Scene 1

I'm trying to picture a world where most courses of study are MOOCs. Universities would almost certainly have far different shapes. Many or even most classes might need no local teaching whatsoever. Local professors or teachers would lose most of their autonomy and become merely assistants to the MOOC prof, leading recitations, labs and group discussion sections. If so, what would become of the Research mission of the University? Would it become even more focused in a tiny number of major research universities?

Would manpower intensive courses like music, art and theater become the province of specialty schools?

The crystal ball is very cloudy, but vast change appears to be coming.

Peer Grading

Much of the magic of the Massive Online Open Course (MOOC) system comes from the professors robotic assistants: the software that keeps track of your course work and assignments and gives you instant feedback on quizzes and exams. What about essay questions? MOOC software can grade short answer questions, albeit a bit imperfectly, but I have no faith in its ability to deal with the longer form. For that reason, MOOCs have been experimenting with peer grading. I'm not sure how it's going, but I'm not optimistic.

Of course scholars and scientists are hardly strangers to peer review. It sort of works for science, but reviewers assigned are usually supposed to be experts in the field of the paper, and in any case have all accomplished earning some of the checkmarks of academic success - something hardly likely to be the case in your freshman composition course.

I suppose I will have to take one of those peer graded courses, but I'm not looking forward to either the giving or receiving end of peer reviews. Of course I don't much like reviewing others scientific papers either. It's hard and usually thankless work.

Monday, May 27, 2013

Arguments With Humanities Professors

I have been debating the merits of the new Massive Online Open Courses (MOOCs) lately, mostly with a bunch of humanities profs. It would be more precise to say that they come up with all sorts of reasons MOOCs can't, and shouldn't be allowed to work or exist. Essentially they argue that the in person prof is the key to actual learning. It's not obvious that this is false.

Why do we need profs anyway? To design the course? A book can do that, and can be more detailed and more precise. To evaluate student performance? Well maybe, but robo-grading works for factual and technique testing, though maybe not for grading essays - yet anyway. To give lectures? They are better on my computer, I think, where I can pause them when I feel like it, or repeat segments when I need to. To answer questions? OK, but Wikipedia can do that. I seem to be left with the task of leading and guiding class discussions.

Am I missing anything?

Sunday, May 26, 2013

Michael Kinsley is a Colossal Jerk (as well as an Idiot)

Pretty much everybody has weighed in on Michael Kinsley's offensive and spectacularly ignorant column on austerity - see Krugman and DeLong for some of the detailed follies, but I like Jason Linkins exposure of its contemptible immorality:

Kinsley: "The problem is the great, deluded middle class -- subsidized by government and coddled by politicians. In other words, they are you and me. If you make less than $250,000 a year, Obama has assured us, you are officially entitled to feel put-upon and resentful. And to be immune from further imposition."

Linkins points out that it is exactly this group that has paid, not for their own sins, but those of the megacapitalists and bankers who created the great recession.

Linkins:"The group of people who Kinsley refers to as the "great, deluded middle class" happen to be the people who have already paid a huge price for the profilgacy of those who personally took actions that specifically tipped the economy into a downturn. And there's a dollar figure you can put on the price they paid: $4.7 trillion. That's how much taxpayer wealth was "disbursed by the U.S. government in an effort to aid the financial services industry," after the financial services industry cocked up the economy of the whole damn planet. The middle-class has also paid for that error by being subjected to a massive unemployment crisis."

I don't take his numbers at face value, but the notion is certainly correct

Thursday, May 23, 2013

Race as a Social Construct

Andrew Sullivan and Ta-Nahisi are thumping away at that old meme of "race as a social construct." Well, all our concepts are, you know. One point that is absolutely clear is that there are no magic lines which one can use to divide humans neatly into three, five, or twenty-seven clear and distinct races. At the same time, most of us can look at another of us and guess something about where most of our ancestors came from.

Populations differ on average, and differed more before the modern era started moving people around in large numbers. It's perfectly obvious that there are systematic differences, some of them easy to recognize, between populations that have lived in Europe or the Congo for six hundred years. Some of those differences will persist in those with some ancestry from one place or another that that have lived in another place for a couple of hundred years.

The fact that race is a social construct, with no absolute biological undepinning does not say anything about the fact that, for example, Americans with more African ancestry might or might not have more musical talent, or running speed, or various other attributes. Of course the ugly cousin in the corner is the nasty issue of IQ. I'm pretty sure that we don't know enough right now to say whether the differences in IQ noticed between populations are of social, genetic, or other origin. Eventually we will know something about the biology of IQ.

In the meantime, attempts at social engineering, say via. immigration policies that discriminate by race, are as pointless as they are invidious.

What is Watson's IQ?

Watson, the IBM Jeopardy champion and medical diagnosis expert, has shown that it can beat humans at a bunch of intellectual tasks. How about IQ tests? Probably IBM is too smart to tackle that problem, but my guess is that some amateur with a punier computer will do it in the not too distant future. Raven's progressive matrices, look out!

If nothing else that might make the race and IQ debate more irrelevant.

Monday, May 20, 2013

Brave New Genome

Something like 3500 Mendelian (single gene) genetic diseases have been discovered but the vast majority of our genetic diseases appear to be multi-genic. Something like 90 genes are known to be associated with Schizophrenia. Some of these involve proteins that have functions that can be directly linked to the brain. Four of them code subunits of a calcium channel protein, that is, a protein involved in the flow of ions into and out of cells, known to be active in nerve cells.

Elucidating the exact way each of these contribute to disease will take a while, but diseases are not the only area where modern genomic statistical analysis will elucidate genetic influence. Height, running speed, and adult IQ are all other areas known or strongly suspected to be under the control of multiple genes. Genomic analysis has developed powerful tools for discovering the function of identified genes, and investigators have gotten rather good at designing customized mice. The designer human will almost certainly be within reach in one or at most two decades.

This will present a temptation that will be difficult to resist. In the first place parents will want offspring that don't have any of the obvious disease causing genes. Once you start tampering with the genome, can you deny your child a slightly more athletic build, a prettier face, or a few more IQ points?

Sunday, May 19, 2013

The Old Ones

It's a commonplace but rarely appreciated fact that our planet is being managed mainly for the convenience of a bunch of billion year-olds. These ancient ones have been around almost since the formation of the planet. They change very slowly, just a few little tweaks every million years or so. They are parasites of a sort, carrying on their business in fish, trees, people and even bacteria.

Of course if they were the sort to talk, they might say it differently. They might say that the bacteria, trees, and people are just convenient dwelling spaces they have devised for their accomodation. I'm writing, of course, about the society of DNA molecules, each of them a clone of a clone and descended in unbroken lineage from some first DNA molecule 3.5 billion years or so ago. With DNA clones though, there is little distinction between the copy and the original, since each new molecule incorporates half of the original and is a copy of it, so in effect each of those DNAs today *is* that original molecule.

I find it more than a little humbling to think that all our works - our art, our science, and our civilization are just the latest trick these little rascals have devised for their propagation.

They don't plan ahead, of course, and that could be their downfall. They have just one central trick, carrying around the instructions for constructing all the stuff they need for their survival. But now their latest invention has devised other kinds of memories and instruction sets, ones that could be more durable and versatile.

The ancient ones, too, may ultimately need to bow to our robot overlords.

Saturday, May 18, 2013

Terror and Identity

The Boston marathon murderer allegedly left a message in the boat where he hid saying that his attack was in response to the war in Afghanistan. "An attack on any Muslim is an attack on all," he proclaimed. This notion of group solidarity is at the heart of the idea of gang, tribe, clan, and probably religion as well. This statement has sort of mirror image - "an attack by any Muslim is an attack by all."

The logical conclusion of this reasoning is war until the extermination of one group or another - or perhaps all groups but one. The evidence suggests that throughout most of human history war has been a major cause of human mortality, and that we evolved to compete in small groups with each other. That instinct to organize against each other is at the center of racial, religious, and national strife as well as the extremes of gang warfare and political partisanship.

If you sell an LED light as cheaper than its incandescent counterpart, both conservatives and liberals are more likely to buy it. If you add that it also helps conserve our resources liberals like it better and conservative like it less. If you give some teen aged summer campers red baseball caps and the rest blue ones, they will quickly evolve into warring tribes.

Civilization depends on being able to tame the tribal (or clan) monster within. The trouble is that it doesn't take very many Tsarnaevs too set off a fatal religious war.

Wednesday, May 15, 2013

MOOC News

a "great dark wave climbing over the green lands and above the hills, and coming on, darkness unescapable. I often dream of it." Faramir to Eowyn in the LOTR.

Those who see the coming of the MOOC as the stroke of doom might see that dark wave in Georgia Tech's announcement of a planned online Masters degree in Computer Science, delivered by MOOC, via Udacity. Now online degrees aren't a novelty. Arizona State, a couple of notches below GA Tech on the academic heap, offers some 60 of them, including an MS in electrical engineering, and the for profit University of Phoenix has been in the business for a few decades.

What is new in the MOOC world are the economies of scale stemming from a lot of the work being done by robots. The Ga Tech MS is supposed to be priced under $7000. For comparison, by my computation, tuition alone for ASU's BA in Art History (now there's a marketable skill) would run about $60,000.

Other people who have to be seeing the competition gaining on them have to be Coursera and edX. The founders of Udacity are hard chargers from Google and Stanford, so they need to be taken very seriously. In the meantime, how about schools that offer nothing but the priviledge of being exposed to their faculty? Yale and Princeton are too stingy to even give you an acknowledgement of course attendance, much less a certificate.

Put up or shut up time is coming to higher ed.

Tuesday, May 14, 2013

Data Wars: Climate

I recently got to argue with a bunch of old friends and former colleagues about climate. It seems that they had formed a little group of climate warriors who would meet for breakfast and the bashing of the conventional climate wisdom. One point of physics they were confused on was the surface temperature of Venus. the guru of their group, a retired meteorologist and meteorological researcher, was convince that the thickness of the atmosphere and the gas law alone could explain the temperature at the surface. Of course they also didn't know that Venus, despite being only 2/3 of our distance from the Sun, absorbed less solar radiation than Earth.

Heat is transported in the atmosphere through three mechanisms: conduction, convection, and radiation. Most of the heat energy we get from the sun is absorbed at the surface, and since we know that the planet has been at roughly the same temperature for billions of years, essentially all of that absorbed energy is returned to space. Space is too empty to carry away heat by conduction or convection, so that all that energy the planet absorbs is reradiated into space. The absorption of solar energy, as I mentioned, takes place at the surface, but most of the reradiated energy is emitted from a region near the top of the troposphere. That fact, and the processes that mediate it, is responsible for greenhouse warming. If the atmosphere were transparent to infrared radiation the re-radiation would take place from the surface, and the planet would be a lot colder than it is.

The lower part of the Earth’s atmosphere, the troposphere, has a temperature structure that is dominated by convective processes. That fact is due to it being relatively opaque to infrared radiation – when radiation transport is efficient it dominates convection since light moves a lot faster than molecules. The dominance of convection produces the well known decreasing temperature with altitude, with a rate of decrease that tends to approach but only locally exceeds the adiabatic lapse rate – that is the lapse rate achieved by transporting a parcel of atmosphere adiabatically to higher altitude.

It’s this low atmosphere dominance of convection over radiation that creates the troposphere, and when the density of absorbing gases gets small enough, radiation dominates and convection halts. Note that putting more absorbing gases into the atmosphere tends to increase the depth of the troposphere, since it increases the height at which radiation dominates.

Once we appreciate this point, we can see why Venus has such a thick troposphere. Of course it is the thick atmosphere, about 90 times as massive as that of Earth that makes it possible, but it is the opacity due to that thick atmosphere which is responsible for the thickness of the convecting zone. An equally massive atmosphere of much smaller opacity would lead to a lower tropopause and surface temperature.

Friday, May 10, 2013

A Sad Transition

Boy genius to middle-aged nut job. Sad but true.

Tuesday, May 07, 2013

Credential

Credential, meaning something signifying worthy of faith or belief, is a word with a distinguished heritage, likely derived from the Proto-Indo-European word for "heart", "to believe," or "put one's heart." As everyone knows, except perhaps a few liberal arts profs, awarding credentials is one of the primary businesses of higher and other education. Students, parents, and legislatures put their faith, money, and heart into education so that their children and society generally, will be enriched by achieving these credentials.

Coursera, the biggest MOOC, has just made a major move into continuing education for educators. This is an opportune time, I suspect, since they hope to take advantage of the big move to the Common Core standard. Teachers are one of the principal consumers of continuing education, so this is a big business area especially for State universities. Teachers are big consumers of continuing education since they pay is closely tied to post graduate education. If school districts and States will sanction Coursera's classes for teacher credit, teachers will flock to them. Expect a big push in that direction, and a big pushback from schools afraid of being passed by.

Friday, May 03, 2013

High Stakes Testing in Higher Ed

George Bush and No Child Left Behind dragged elementary and secondary education into the world of high stakes testing, albeit kicking and screaming. What about that really pricey item in the budget, higher education? So far it has been mostly immune, though there are a few clues from such things as the GRE, MCAT, LSAT and so on, but are there any institutional statistics kept for those things? I haven't heard of them, at least not of public statistics.

The country spends a fortune on higher ed, so wouldn't it make sense to see what we are getting for our investment? I think that the Obama administration wants to see something along these lines, especially to rein in some of the outrageous nonsense being sold by the for profit colleges.

If you are in higher ed, though, you have to be at least a little worried. Suppose you are Harvard, with astronomical tution, entering SAT scores, and resources, and it turns out that your graduates really aren't outdoing those of State U X? That would be a prestige calamity of catastrophic proportions. What about Nowhere State U, if it finds out that its grads, with comparable SAT scores to those at Barely Anywhere State are losing by a mile? And what about those pesky MOOCs?

Implementation presents challenges. Who gets to decide what every graduate ought to know? How about majors and concentrations? For physics, math and engineering, this shouldn't be too tough, but what about philosophy, history, and theater?

Whatever the test, it should not be a minimum proficiency test. Make some problems easy enough even for those who spent most of their six years in school stoned, and some hard enough to challenge the prodigy.

Panic in University Park

Fear has descended on the academy. Many are starting to see the threat posed by the MOOC. Of course higher ed, or at least the more vulnerable reaches of higher ed, haven't been happy for a while. Having tried insults and related tactics, those who perceive themselves threatened are now trying to appear to the compassion and brotherly feeling of their fellow profs who man the MOOCs. Here is a sample, courtesy of Jonathan Rees, the proprietor of a blog lately devoted mostly to anti-MOOC activism: More or Less Bunk.

I know I’m late to the party on this, but that letter to Harvard’s Michael Sandel from the San Jose State (SJSU) Philosophy Department really is quite wonderful. I’m going to try to take up its implications with respect to academic freedom and shared governance over at the Academe blog as soon as I get my grading done, but what I want to discuss here is the way that those nice folks in California actually called out Sandel, not just their administrators.

You can see this most clearly at the very end of the letter:

“We respect your desire to expand opportunities for higher education to audiences that do not now have the chance to interact with new ideas. We are very cognizant of your long and distinguished record of scholarship and teaching in the areas of political philosophy and ethics. It is in a spirit of respect and collegiality that we are urging you, and all professors involved with the sale and promotion of edX-style courses, not to take away from students in public universities the opportunity for an education beyond mere jobs training. Professors who care about public education should not produce products that will replace professors, dismantle departments, and provide a diminished education for students in public universities.

This appeal stings, to be sure, but will it work? The trouble is, the protesting professors are asking the MOOC teachers to put the economic interests of their fellow professors ahead of their students. Of course that's not the way teh reactionaries would put it, but if you believe in what you are teaching at your MOOC, how can you think otherwise.

Of course it's possible that the MOOC is really only suitable for a special class of highly motivated and self-confident students who would disdain professorial hand-holding anyway. Or maybe they only work in technical classes. At some point, perhaps even soon, some post class testing is going to be introduced, and then MOOCs and professors alike will find their work under the eye of the educational panopticon - probably even more that elementary and secondary teachers already are.