What are you going to simulate?

The EU is sinking €1.2bn (and the US is proposing to spend more, $3 billion) into a colossal project to build a supercomputer simulation of the human brain. To which I say, “What the hell? We aren’t even close to building such a thing for a fruit fly brain, and you want to do that for an even more massive and poorly mapped structure? Madness!” It turns out that I’m not the only one thinking this way: European scientists are exasperated with the project.

"The main apparent goal of building the capacity to construct a larger-scale simulation of the human brain is radically premature," Peter Dayan, director of the computational neuroscience unit at UCL, told the Guardian.

"We are left with a project that can’t but fail from a scientific perspective. It is a waste of money, it will suck out funds from valuable neuroscience research, and would leave the public, who fund this work, justifiably upset," he said.

There is a place for Big Science. I’d suggest that when you’re at the preliminary exploratory stage, as we are with human brain function, it’s better to fund many small exploratory parties to map out the terrain, rather than launching a huge invasion with charts that are made out of speculation. We know a computer simulation is going to fail, because we don’t know what it’s going to simulate. So why are they doing this? Maybe it’s a question of who “they” are.

Alexandre Pouget, a signatory of the letter at Geneva University, said that while simulations were valuable, they would not be enough to explain how the brain works. "There is a danger that Europe thinks it is investing in a big neuroscience project here, but it’s not. It’s an IT project," he said. "They need to widen the scope and take advantage of the expertise we have in neuroscience. It’s not too late. We can fix it. It’s up to Europe to make the right decision."

I’ve noticed this, that a lot of gung-ho futurists and computer scientist types have this very naive vision of how the brain works — it’s just another computer. We can build those. Build a big enough computer, and it’ll be just like the brain. Nope. That’s operating on ignorance. And handing ignorant people billions of dollars to implement a glorious model of their ignorance is an exercise in futility.

Closure on the Obokata/STAP affair

I’ve been following the story of stimulus-triggered acquisition of pluripotency (STAP) cells with considerable interest, and there’s a good reason for that: from the very beginning, it contradicted how I’d always thought about cell states, and if it were true, I’d have to rethink a lot of things, which was vexing. But on the other hand, empirical results always trump mental models, so if the results held up, there was no question but that I’d have to go through that uncomfortable process of reorganizing my preconceptions. It would be OK, though, because there’d be a great prize at the end.

Well, it turns out that I don’t have to reboot my brain after all, because now that all the flailing about is over, STAP is a product of sloppiness and fakery, and is dead.

So here’s the controversy, and why I found it vexatious. We want to be able to specify cell states; in particular, we’d love to be able to take any cell from the human body, tickle it with a few specific signals, and see it throw away all of its historical constraints and become a different cell type altogether. In particular, the Holy Grail is to find the right combination of switches to cause any cell to become a pluripotent stem cell — the kind of cell we can then induce to become any other cell type we might need.

We know this can’t be impossible, and is probably even fairly simple, because we know that cells can do this already (well, to some degree; your body accomplishes this task by setting aside reserve populations of stem cells. It’s also likely that some cell types are so tightly locked in by the process of differentiation that their state is not reversible). The idea is that we just need to find the right combination of signals/genes — the right kind of key — and we can unlock the cell, and make it open to additional inductions that will allow us to manipulate it.

We have some idea of the shape of the key. Yamanaka identified four genes, Oct4, Sox2, cMyc, and Klf4, that when activated, switched cells into a pluripotent state, making induced pluripotent stem cells, or iPS cells. It works. The handicap right now is that we only have a kind of brute force method of switching those genes on, and two of them are oncogenic, so it’s as if we’ve got a rather clumsy key that opens the lock, but also damages it in unfortunate ways. The resolution to that problem, though, was learning how to finesse the genes — we need to figure out how to more delicately switch on the necessary genes by a way other than bluntly transfecting cells with copies of the genes that are always on.

Then along came Haruko Obokata, an investigator in Japan who announced that she could induce stem cells with simple, generic stress, such as by exposing them to acid or physically pushing on the cells. It was like saying she didn’t need a specific key, all you needed to do was shake the lock really hard, and it would spontaneously pop open. What, really? That just seems too simple. It would be phenomenally awesome if true, but it seemed unlikely. But then, I remember this one lab I worked in where all the publicly popular drugs, like ketamine, were kept locked in a drawer to which only the PI had a key…but the countertop wasn’t secured to the bench, so if you knew about it, you could just lift the top and get easy access. It was a backdoor to the goodies that was so stupid you couldn’t believe it existed, but it did.

Could it be that cells similarly had a stupid weakness that could be so easily exploited? The short answer is no; read the whole article by David Cyranoski.

But the paper1 that set out the fundamental technique was soon shot full of holes. There was plagiarized text in the article. Figures showed signs of manipulation, and some images were identical or nearly identical to those used later in the same paper and elsewhere to represent different experiments. More damning were genetic analyses that strongly suggested the cells were not what they were purported to be. And although deriving STAP cells was advertised as simple and straightforward, no one has yet been able to repeat the experiment.

Within the space of six months, Obokata was found guilty of misconduct by her institution; well-respected scientists, including RIKEN head Ryoji Noyori, bowed their heads in apology; and both papers were retracted. In the end, the evidence for STAP cells seemed so flimsy that observers began to ask where were the extra precautions and the ‘extraordinary proof’ that had been promised post-Hwang.

It sure would have been nice to have a simple technique for generating stem cells, but I have to confess to being a bit relieved. There’s the vindication of prior thinking and the value of incrementally improving our stem cell protocols, of course, but also, I’d personally rather that it weren’t trivial to switch my cells to a de-differentiated pluripotent state — that’s a recipe for easy cancer generation, too. It is somehow reassuring to think that evolution has shaped multi-cellular organisms to be somewhat resistant to spontaneously going all stem-celly under stress.

Terrible disease rips through BBC staff

The news out of the UK is grim. Various voices in the media are falling silent, victims of an affliction called “reason”. The staff have been told that false impartiality, which allows kooks to air their views side-by-side with legitimate experts, must stop.

BBC journalists are being sent on courses to stop them inviting so many cranks onto programmes to air ‘marginal views’

The BBC Trust on Thursday published a progress report into the corporation’s science coverage which was criticised in 2012 for giving too much air-time to critics who oppose non-contentious issues.

The report found that there was still an ‘over-rigid application of editorial guidelines on impartiality’ which sought to give the ‘other side’ of the argument, even if that viewpoint was widely dismissed.

Some 200 staff have already attended seminars and workshops and more will be invited on courses in the coming months to stop them giving ‘undue attention to marginal opinion.’

They specifically mention anti-vaccine kooks, climate change denialists, and GMO hysterics, but I imagine it applies to creationists, flat-earthers, and people who claim to be able to square circles. For the BBC, this disease is going to sweep through them like a high fever requiring a bit of bed rest — they’re going to have to kick interviews with James Delingpole or Christopher Monckton to the curb.

But if Reason proves infectious and jumps the Atlantic, sweeping through American newsrooms, the effects could be devastating. We have no natural immunity. Our media revels in crankery of all kinds. Imagine this rule enforced on the executives of the History Channel: we’d have 24 hours of dead air. What if Fox News came down with it? It’d be like the Zombie Apocalypse there. Roger Ailes would have to be hospitalized; Fox & Friends would be populated with stunned, broken, speechless idiots staring teary-eyed and mute at one other; Bill O’Reilly wouldn’t be able to vent gas and would eventually explode. The Sunday morning pundit shows across all the networks would be destroyed. Imagine if they had to face the fact that Dick Cheney was disastrously wrong and simply not a respectable source to be consulted on foreign affairs?

Oh, the humanity.

I will not call on you to demand Anthony Cumia be fired

Anthony Cumia of the Opie & Anthony show has a long history of public awfulness. He’s a sexist pig and a creep.

But I don’t think you should call Sirius XM and complain.

His latest episode was a flamingly racist tirade against a black woman (only he didn’t restrain himself to merely call her a “black woman”). He wanted to shoot her because she slapped his camera away when he was taking creepshots.

He’s an appalling human being. But why bother demanding his dismissal?

Many media outlets are howling about his violent racist fantasies. He’s scum.

But he’s just the erupting pimple of the problem. The real issue is that somewhere in the corporate headquarters for Sirius XM, there is a nest of verminous, amoral, soulless corporate drones who saw a racist misogynist loudmouth as a pile of dollar signs. Fire Anthony Cumia, they’ll still be there. Fire Cumia, his audience of sympathetic racist misogynist cowards will still be there. Treating the repugnant excrescences without digging deep to the root of the disease is not enough.

If you want to do anything, cancel your Sirius XM account. Not conditionally, not if they don’t fire Cumia, but just plainly and simply cut them off. Punish the executives. Do you know anyone who listens to that Opie & Anthony crap? Repudiate them, publicly and unabashedly. Let them know that they are also terrible human beings for giving an audience to racists.

Firing Anthony Cumia is just the icing on the cake. Demand more.

If your only justifications for sexism are stupid, you should maybe stop making them

Once again, we get stupid answers to a good question. A a guest post on the Curious Wavefunction decides that Larry Summers was right, there are innate differences between men and women. (Curiously, this is the same blog that posted a positive review of Nicholas Wade’s book — strange how sympathy for racism and sexism go hand in hand). It starts off well by pointing out a real phenomenon, the different sex ratios found in different scientific disciplines.

Here are statistics on the sex ratio among graduate students. The order here is by level of analysis (with computer science thrown in somewhat arbitrarily next to physics). The first number is men; the second number is women.

Physics:   1694: 448
Computer Science:  1465: 380
Chemistry:  1520: 897
Biology:  3936:4494
Psychology   1047:2566
Anthropology:  186: 360:
Sociology: 230: 400
Political Science: 422: 303

Why is that, you might wonder…and of course, the answer we’re going to get is that there’s something about the Y chromosome or the hormonal environment that predisposes one to like computers vs. cells, or experiments with lab rats vs. electronic gadgets. Which, simply on the face of it, is complete bullshit.

I’ve noticed that these arguments often prefer to show a current snapshot of the statistics to make their point: looking at historical trends tends to screw up their assertion of a biological difference. A century ago, virtually none of these disciplines had a preponderance of women enrolled in them, and women professors were extremely rare. I guess there has been a remarkable degree of selection extinguishing all those stupid women from the population in the last 100 years. It’s simply not possible that cultural factors might strongly influence the pattern.

Similarly, we could ask questions about aptitude. Women just don’t like those disciplines with low female enrollment, because ladybrains. But how do we account for historical shifts in ability? Look what’s going on in the British school system.

The relative improvement in girls’ performance in examinations at 16 has been achieved over the last ten years. In the l960s, boys outperformed girls by about 5%; for the next fifteen years, boys and girls were performing at almost equivalent levels. However, from 1987 only about 80 boys to every hundred girls achieved 5 high grade passes at 16+. Boys lost their advantage in terms of school leaving credentials and are now struggling to keep up to girls’ success rate. In the mid l980s, girls turned the tide of credentialism, even at least temporarily, in their favour.

Oddly, no one seems willing to advance the bold hypothesis that maybe boys’ genes and hormones make them less scientific. It’s always the other way around, that women are less capable, because ladyparts…even in defiance of the evidence that women are performing better than men.

OK, well I promised we’d have some good questions. I lied. This guy asks stupid questions.

This brings us to two related questions: Why is the percentage of women somewhat proportional to the “socialness” of the science?

WHAT THE HELL…? Look at the list up above. Can you tell me which of those disciplines is more “social” than the others? Science in every discipline is an extremely social enterprise — if you’re going to succeed in it, you have to be able to engage with your colleagues, present your work publicly, collaborate, teach, and work in committees. If you think you can crawl into a basement and do computer science without bathing for 5 years, well, you can…but you won’t get a job afterwards, and you won’t be able to be a significant team member. Really, I know computer scientists. They do bathe regularly, and they can be friendly and engaging.

This assumption is a classic example of circular reasoning. Women are more social; some disciplines have more women than men; therefore, biology must be more social than physics; and the evidence for that is that biology has a higher proportion of women graduate students than physics. Ta-daa. QED. Guys win.

And why don’t women choose academic careers after they finish graduate school? To answer these questions, it’s worth looking at Steven Pinker’s contribution to the post-Larry Summers debate at Harvard.

Wait. That’s a different question, and it’s informative that the distinction isn’t addressed. Women are succeeding in greater numbers in graduate school than they are in post-doctoral careers. That’s revealing! By getting through graduate school, which is a pain-in-the-ass and a major sacrifice of time and money, these women have already shown a strong interest and ability in science. They are full-on dedicated scientists.

What’s the difference in the transition from life as a graduate student to the professoriate? More expectations for teaching, networking, committee work, grant writing, and collaboration. More of those “social” skills we’ve just been told women are naturally better at. As usual, none of this argument makes any sense, and consists entirely of sloppy attempts to rationalize prior biases.

We are promised evidence that women are simply less suited to working in science. Let’s see the list.

The full debate is also worth reading in full—and I apologize for giving Elizabeth Spelke, Pinker’s opponent, short shrift here–but this is Pinker’s summary of the psychological differences between men and women:
1. Men, on average, prioritize status, while women weigh status and family equally.
2. Women, on average, are more interested in people; men are more interested in things and abstract rule systems. 
3. Men are by far the more reckless sex.
4. Men, on average, have a superior ability to do three-dimensional mental transformations.
5. Men, on average, are superior at mathematical reasoning.
6. Men have more variability than women across traits, which means that men are over-represented in the upper and lower tails of ability distributions.

Jebus. Those are just assertions. Who says women don’t “prioritize” status? How was that measured? Do women just sit around content to be egalitarian? Has anyone considered the possibility that status-seeking is going to be entirely culture dependent, and that there are different ways for different people to achieve high status? (I know they have, but those complexities are always jettisoned by the advocates for male superiority — the parameters of male dominance are naturally and obviously the only ones that matter, and they are objectively independent of societal constraints.)

I’m not going to was time going through these bald assertions one by one, but let me just say, they don’t apply. We’re already looking at a fairly rarefied subset of human endeavor occupied by people of largely above average socio-economic status, from families with an already above average emphasis on education. You can’t derive the properties of an already select subset from generalizations about the population as a whole, or you’d have to conclude that most scientists are blue-collar and service workers. We already know that there are strong cultural and familial influences that bias some women to pursue careers in science — factors that are not present for most women. Or most men. So even if their generalizations about the abilities of women were valid on the whole (and I don’t agree that they are; #2, 4, and 5 are clearly influenced by social conditioning, and #3 is statistically true but again probably influenced by social expectations), you simply can’t use them to talk about a small population that is both self-selected and strongly constrained by socioeconomic opportunities.

But #6…oh, #6, how I despise you, and how I constantly hear it trotted out as a justification. Look at the structure of that argument. We can clearly show that more men than women have mental disabilities and illness, therefore, we should expect that more men should show a greater range of high intelligence.

WHY? That makes no sense. It’s some kind of weird appeal to statistical fairness — that for some reason, Nature must balance every curse with a blessing, that every curve must be a perfectly symmetrical bell shape. But that is not true. There is no reason to expect it to be true.

It’s also selectively applied. I noted up top that the same people who argue for differences in intellectual aptitude between the sexes also like to argue for differences in intellectual aptitude between races. Yet for some reason, you’ll never see them suggest that since black people have a higher frequency of criminals and ignorant, uneducated men (I’d suggest that poverty and discrimination play a strong role in that, but you know these guys — they say it’s objective evidence of genetic inferiority), they must therefore also have a greater proportion of saints and geniuses in their populations.

It would be only fair, you know. Everything has to balance.

Or we never see this interesting proposal: malnutrition also increases the variability in a population. Therefore, if we want more supergeniuses, we should starve children — sure, we’d get a lot of death and illness, but the ones who thrive are going to be tough and brilliant.

I find myself endlessly exasperated by these transparently stupid justifications. They aren’t even internally consistent.