Category Archives: Mathematics

AI In Medicine

Self-taught systems beat MDs at predicting heart attacks:

All four AI methods performed significantly better than the ACC/AHA guidelines. Using a statistic called AUC (in which a score of 1.0 signifies 100% accuracy), the ACC/AHA guidelines hit 0.728. The four new methods ranged from 0.745 to 0.764, Weng’s team reports this month in PLOS ONE. The best one—neural networks—correctly predicted 7.6% more events than the ACC/AHA method, and it raised 1.6% fewer false alarms. In the test sample of about 83,000 records, that amounts to 355 additional patients whose lives could have been saved. That’s because prediction often leads to prevention, Weng says, through cholesterol-lowering medication or changes in diet.

To be honest, while it’s statistically significant, I’d have expected a bigger improvement than that. And it’s not clear how useful it is if the recommendations aren’t science based, as prescribing cholesterol-reduction or diet change generally aren’t.

The National Debt

The budget cuts don’t go deep enough:

The chart is brutally bipartisan. Debt increased under Republican presidents and Democratic presidents. It increased under Democratic congresses and Republican congresses. In war and in peace, in boom times and in busts, after tax hikes and tax cuts, the Potomac flowed ever deeper with red ink.

Our leaders like to talk about sustainability. Forget sustainable — how is this sane?

Yet when any politician hesitates before increasing spending, he’s portrayed as a madman. When Paul Ryan, R–Wis., offered a thoughtful plan to reduce the debt over decades, he was pushing grannies into the Grand Canyon and pantsing park rangers on the way out.

One of the many cons of Trump is that his budget cuts are going to solve the problem, when they don’t even scratch the surface. I’m not opposed to them, in general, mostly because I think they’re funding things that aren’t a federal responsibility, but it’s ludicrous to think they’ll have any significant fiscal effect. As long as he refuses to touch entitlements, we’re fiscally doomed.

Trump’s Climate EO

Ths hysteria on this from the Left has been some combination of frightening and hilarious.

But in fact, as Roy Spencer points out, the (illegal) “Clean Power Plan” was literally going to increase poverty and kill people, while almost certainly having no discernible effect on climate.

For Deluded Warmists

A handy primer:

CO2 levels were steady during these wild swings and throughout the Holocene at roughly 280 parts per million (ppm) until 130 years ago when a stuttering increase to 400 ppm today began. In other words, Holocene temperature changes, and the wild variations that preceded them, were not linked to CO2 changes. This prompts the question: if CO2 changes did not drive these temperature shifts, why all the fuss about CO2 emissions?

The answer owes much to the complexity of the climate system and the wish for simple explanations to explain its variability and with which to make predictions. But climate is not simple. There are many interacting parts that make it a ‘coupled non-linear chaotic system’ in which small variations of any part can create big, unpredictable changes. In the search for something simple to blame, like increasing CO2 levels, this ‘coupled non-linear, chaotic’ nature of climate is often played-down, overlooked or ignored. Things like solar variations, ocean heat transfers, cloud cover and the like – things that may well be the main drivers of climate – seldom get the respect they deserve.

The effect of the sun, the sea and clouds on climate is known and accepted – the Gulf Stream being a well known example – but more precise knowledge suitable for computer models is a different thing altogether. But what can be said for sure, is that the sun, the sea and the clouds are all very important and CO2 is only one player in a big game, not the control knob on the Earth’s thermostat. It is true that CO2 contributes to the greenhouse effect, but its heating effect is small (when compared with water vapour, the main contributor) and drops off logarithmically as its concentration increases. The more there is, the less additional heating effect it has.

It’s almost as though there’s some sort of political agenda that has nothing to do with science or reality.

[Update a while later]

This is interesting: Pruitt doesn’t want to attempt to overturn the endangerment finding:

Pruitt, with the backing of several White House aides, argued in closed-door meetings that the legal hurdles to overturning the finding were massive, and the administration would be setting itself up for a lengthy court battle.

A cadre of conservative climate skeptics are fuming about the decision — expressing their concern to Trump administration officials and arguing Pruitt is setting himself up to run for governor or the Senate. They hope the White House, perhaps senior adviser Stephen Bannon, will intervene and encourage the president to overturn the endangerment finding.

Trump administration officials have not totally ruled out eventually targeting the endangerment finding. Conservative groups have petitioned the EPA to look at reopening it, one source said, and the agency may eventually be compelled to respond to the petition. Axios first reported the news of the petition.

“Getting rid of the Clean Power Plan is just not enough,” said Myron Ebell, the director of the Center for Energy and Environment at the Competitive Enterprise Institute and the former leader of Trump’s EPA transition team.

I agree. It was based on junk science. In fact, they should be trying to get a rehearing of Massachusetts versus EPA when they get Gorsuch on the court.

Science And Policy

Words of wisdom from Daniel Sarewitz:

Whatever science you’re doing on a post-normal problem, it is always going to be incomplete, and it is always going to be subject to revision, and highly uncertain. It can be viewed from numerous scientific perspectives. So multiple scientific studies can come up with multiple results, so it leads to a profusion of truths that can be mobilized on behalf of different sets of values. Values and facts can pair up with each other in different ways.

One example I love is how everyone talks about how there’s a consensus on GMOs. Well there is consensus around a narrow part of the GMO issue, like there is a consensus around a narrow part of climate change. But the real problems have to do with the ‚what could be done?‘ questions. So for GMOs for example, when people say there is a consensus, what they mean is ‚we know they’re not a health risk‘. So I’ll accept it on health risk, I don’t have a problem with it. But then you say, ‚and we know that they’ll be an essential part of the economic future of Africa‘. Well, maybe that’s true — whose model are you using? What kind of data have you used to generate that? What are your assumptions? I mean anything dealing with projections of the future and claims about how the world is going to look, in a multi-variate, open system, are going to be subject to different people coming up with different claims and conclusions. And that’s exactly what happens.

And when you bring science into the political debate, you have to pick and choose which science you want to use. You have to match that with particular priorities about what policy problems you want to solve. I think science is really important, I think we want to be factual, I think we want to have a grip on reality and I think science can help us do that. But for problems where there are so many paths forward, so many competing values, the systems themselves are so complicated, I don’t think science is a privileged part of the solution.

…The post-normal science idea really does challenge the notion of science as a unitary thing that tells us what to do, PNS really says that we have to think of science in a different way in these contested contexts, and I don’t think most scientists want to go there. The deficit model puts them in charge: “we communicate the facts, you listen and take action.” So if the problem isn’t solved it’s not science’s problem. This is a self-serving superstition that the scientific community generally holds. And superstitions are hard to destabilize.

Over on Twitter, I’ve been having arguments with people about the proposed cut at the EPA, in which the budget for “protecting the climate,” is reduced to “only” $29M.

What in the hell does “protecting the climate” even mean?


wonders why NASA is considering crewing the first flight of SLS/Orion:

In a statement at the beginning of the Feb. 23 meeting of the Aerospace Safety Advisory Panel (ASAP), chairwoman Patricia Sanders said that if NASA decides to put a crew on the first SLS/Orion launch, Exploration Mission 1 (EM-1), it must demonstrate that there is a good reason to accept the higher risks associated with doing so.

“We strongly advise that NASA carefully and cautiously weigh the value proposition for flying crew on EM-1,” she said. “NASA should provide a compelling rationale in terms of benefits gained for accepting additional risk, and fully and transparently acknowledge the tradeoffs being made before deviating from the approach for certifying the Orion/SLS vehicle for manned spaceflight.”

“If the benefits warrant the assumption of additional risk,” she added, “we expect NASA to clearly and openly articulate their decision-making process and rationale.”

The point of my book was not that NASA should simply be more accepting of risk, or be reckless, but balance the risk against the reward. In my opinion, accelerating commercial crew would be worth the risk, to end our dependence on Russia, and increase the productivity of the ISS. Redoing Apollo 8 half a century after the original as a political stunt would not.

[Update a little before 1 PM EST]

NASA is about to have a news conference, probably in response.

[Update post conference]

It was the Bills Gerstenmaier and Hill. Gerst is always deadpan, but one had the impression that he’s not enthusiastic. They’re doing a feasibility study because the White House asked, and won’t be making any recommendations, just describing would it would take in terms of changes in schedule and budget. They just want to see “if they can fly crew sooner.” They expect to have some answers in a month or so (presumably as part of the input for FY2018 budget request). I wish the White House would ask them if they could fly crew sooner on Dragon and Starliner. That would be worth doing.

I can’t believe I just typed the words “FY 2018 budget request.” Makes me feel old.

[Update a few minutes later]

[Update a few minutes later]

Here‘s Keith Cowing’s story.

[Early afternoon update]

And here‘s Eric Berger’s take.

[Update a while later]

And Joel Achenbach’s.

I’d note that the reason they would only have two crew is probably a) to reduce the number of losses if it doesn’t go well and b) more margin in the (primitive?) life support.

[Saturday-afternoon update]

Amy Shira Teitel (like me) thinks that this makes no sense.