SoylentNews is people
SoylentNews
posted by
cmn32480
on Sunday May 08 2016, @11:49PM
from the I-can't-solve-DiffEq-with-a-computer dept.
from the I-can't-solve-DiffEq-with-a-computer dept.
This just in from the front lines of the War on the Unusual:
University of Pennsylvania economics professor Guido Menzio was solving a set of differential equations on a plane departing the Philadelphia airport when the woman next to him surreptitiously passed a note to a flight attendant telling them she thought he was a terrorist because of the strange things he was writing on a pad of paper. The plane returned to the gate where he was questioned. At least this time the pilot had enough sense not to kick him off the flight.
Remember folks, if you see something say something!
This discussion has been archived.
No new comments can be posted.
UPenn Professor Solving DiffEQ on Scratchpad Causes Plane to Return to Gate
|
Log In/Create an Account
| Top
| 107 comments
| Search Discussion
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
But what can you do with it?
-- ubiquitous cry from Linux-user partner
(Score: 2) by devlux on Monday May 09 2016, @12:05AM
What I don't get... If this was a differential equation, isn't most of that Greek symbols?
So even the lame "well he was using Arabic numerals" joke wouldn't seem to apply here, right?
(Score: 0) by Anonymous Coward on Monday May 09 2016, @12:14AM
They thought he was a Muslim but it turns out he's a Numerical Methodist.
(Score: 2, Insightful) by Anonymous Coward on Monday May 09 2016, @12:26AM
You expect the average uneducated american to know the difference between greek and arabic?
(Score: 2) by hemocyanin on Monday May 09 2016, @02:54AM
No, but between paper and a bomb? yes.
(Score: 0) by Anonymous Coward on Monday May 09 2016, @07:34AM
Paper covers bomb.
Bomb smashes scissors.
Scissors cut paper.
None of these are allowed aboard the aircraft. 💣🎫💇
(Score: 2) by c0lo on Monday May 09 2016, @03:02AM
Yeah, naah... it's all greek to me.
https://www.youtube.com/watch?v=aoFiw2jMy-0
(Score: 4, Funny) by frojack on Monday May 09 2016, @01:50AM
Perhaps she knew he was an Economics professor?
They aren't exactly the type you expect to find wielding differential equations.
Much beyond a Chi Square, and they are out of their league.
No, you are mistaken. I've always had this sig.
(Score: 3, Interesting) by opinionated_science on Monday May 09 2016, @03:10AM
I modded you funny, because I had this *exact* conversation with a retired Econ Professor at dinner.
Me: "So do many are many of your theories backed by models with different equations denoting the various economic actors".
Prof: "Well not exactly..."
Me: "So you can see why a lot of us (physicists but you can insert "folks who do loads of maths") think it is all made up".
Prof:" Well to be honest, that's a great deal of it. Emotion and intuition substitute proper model analysis and probability outcomes. So, can you explain why the universe is expanding..."
Verbatim, honest ;-)
(Score: 0) by Anonymous Coward on Monday May 09 2016, @03:16AM
FYI, economics is a branch of psychology, not math. Math comes into play, like it does in practically every discipline, but fundamentally it is about how humans behave under various conditions.
(Score: 5, Informative) by Non Sequor on Monday May 09 2016, @04:34AM
The actual reality is at the same time a bit better than that and a bit worse.
There are differential equation models. They are in wide use. But the simplifying assumptions needed to make them tractable also reduce them to the role of illustrations for different narratives which can be postulated. See https://en.wikipedia.org/wiki/Dynamic_stochastic_general_equilibrium [wikipedia.org]
The problems of statistical inference in economics are much deeper and more fundamental than those in physics. Isolation of variables is impossible in economics. In theoretical physics, in some cases model selection is focused on mathematically minimal objects, no such hope exists for economics. Most macroeconomic variables are heterogenous aggregations (see https://en.wikipedia.org/wiki/Cambridge_capital_controversy) [wikipedia.org] rather than quantities denoted in a single pure unit.
Physicists who model heterogenous systems are called chemists. Chemists who model heterogenous systems are called biologists. Things kind of branch off from there, but you get the point. The lack of your rigor isn't because everyone in the other fields are lightweights (well, economists are lightweights, but that's another thing), but rather it's an effect of a level of aggregation of subsystems underlying each field.
Write your congressman. Tell him he sucks.
(Score: 2, Interesting) by Anonymous Coward on Monday May 09 2016, @08:25AM
Psychologists think they're experimental psychologists.
Experimental psychologists think they're biologists.
Biologists think they're biochemists.
Biochemists think they're chemists.
Chemists think they're physical chemists.
Physical chemists think they're physicists.
Physicists think they're theoretical physicists.
Theoretical physicists think they're mathematicians.
Mathematicians think they're metamathematicians.
Metamathematicians think they're philosophers.
Philosophers think they're gods.
(Score: 2, Informative) by Lester on Monday May 09 2016, @09:08AM
Variable correlation may be casual because there are many other variables, moreover they don't say what's cause and what's effect, or if there are just proxy data. Inference, in economics world, is the final proof and evidence. And that's the problem, inference, in any science, is a good point for formulating an initial hypothesis to start research, but not a fact to build a whole theory on.
If you can't isolate variables, you can't prove things. So it's not science. There are zones of physics that are not demonstrable, contrary to economist, theoretical physicians know it, the economist don't or at least ignore it.
The main problem of accepting as science something that is not, isn't that you can't prove things, but that you can't refute anything. No matter how reality shows results against theory, there is always an unknown, hidden variable that can explain it. There is no progress, no revision, everything is petrified.
When I read "The undercover economist" I thought "What this book proves is that economy at this level is psychology and at nation level is sociology". What a bunch of unproven nosense. i.e. "more abortion, less crime" Excuse me, there are so many variables involved that such statement makes no sense. The problem is not that there are too much variables, the problem is that this guy dared to connect them and write it in a book.
(Score: 2) by Non Sequor on Monday May 09 2016, @11:08AM
So how do you solve that problem without asking people to stop living? Most human decisions are made on matters for which there is no hope of grounding them to clear, demonstrable evidence. The physicist who seeks to apply methodologies that offer clarity in their field and the economist who remains blithely unaware of problems relating to the nature of knowledge both suffer from a form of impotence against the problems of life.
Yet we do know that problem solving strategies exist in these kinds of limited knowledge settings, although no a priori best approach exists (see the no free lunch theorem). This is heuristic reasoning and people use it because, theoretically there are circumstances where it can be effective, and because we've picked an arbitrary strategy out of a range of strategies for which the evidence is not sufficient to identify a clear, overwhelming winner.
Write your congressman. Tell him he sucks.
(Score: 1) by Lester on Monday May 16 2016, @10:45AM
You are completely right. We have to make a lot of decisions with limited data and knowledge.
You can use science or you can use... let's call it intuition. It's amazing in how many cases intuition works well, not only well, but sometimes faster than science. In fact, "What are the scientific facts behind this intuition that seems to work always?" is a common topic of research.
If you can't apply science, because science for that matter doesn't exist, or because a quick acceptable decision now is better than a perfect solution latter, uses intuition. Nevertheless intuition has also a lot of pitfalls and sad record of wrong decisions. So, whenever If you can apply science, apply science. It's the safest bet
So, the first problem is when you insist in using intuition instead of science. And the really big problem is when, instead of using science or intuition, you use bad science or intuition disguised of science.
(Score: 2) by fido_dogstoyevsky on Monday May 09 2016, @10:37AM
The actual reality is at the same time a bit better than that and a bit worse... There are differential equation models. They are in wide use. But the simplifying assumptions needed to make them tractable also reduce them to the role of illustrations...
In other words, the model only works for spherical investors in a vacuum.
2-3 Decades ago a multi millionaire ($1M was a lot of money back then) in Oz made millions by saying he was thinking about using his group of companies to buy some other company out. Couple of weeks later he made some more millions by publicly changing his mind. A small number of major investors makes using statistics difficult.
It's NOT a conspiracy... it's a plot.
(Score: 2) by opinionated_science on Monday May 09 2016, @03:17PM
but those of us that work with statistical mechanics that forms the basis for molecular dynamics simulations, have a giggle at the over-simplistic models.
The point about economics, is that the models are based to a certain degree, upon rational actors.
If there's one thing that we know about the population at large, they do not always act in their best interests (viewed from global point of view).
I suspect strongly, this is why a dart board is at least as a effective as a stock-broker...
(Score: 2) by frojack on Monday May 09 2016, @03:52PM
The point about economics, is that the models are based to a certain degree, upon rational actors.
If there's one thing that we know about the population at large...
Actually, that reliance on rational actors is a myth. If it were that easy you could solve all economic problems sitting in a chair. Economics is all about actively measuring what happens. Actors don't have to be rational, because you don't care about Joe, just Joe Average, and his ilk.
And the assumption that the entire field of economics has anything at all to do with stock brokers is another huge misconception.
Its as much about cheese prices and the flow of money than anything else. All very measurable things. Which is why economists end up being stat-heads. Economics is a "descriptive" science, not an "experimental" one. There is actually very little work done in the way of experimental protocols,
Its more about noticing that when too much milk is produced, one of the ways the market reacts is to store milk for leaner times by making cheese. Governments may want to tinker with the demand for cheese to help dairy farmers, but that's not the economist's fault, any more than it is the Ornithologist's fault that tinkering with the amount of sunflower seeds in a bird feeder can be used to attract more gold finches.
No, you are mistaken. I've always had this sig.
(Score: 3, Informative) by opinionated_science on Monday May 09 2016, @04:45PM
Well my point about "rational actors" is the dogma of low-price , higher volume.
My main criticism of the models used, is the lack of time-dependent effects, and the complete ignorance of entropy (in a closed system).
In molecular systems where we couple microscopic movements to macroscopic properties, we have a configurational integral that defines the partition function of a system.
In economics there are far less stringent couplings, that for example, do not recognise the fundamental limits of human activity - the implicit sampling timestep, that causes inequalities of rational action.
Anyone ever wondered why the news on a sunday is so slow? Or why the markets close at certain times? Obviously this is accommodates human behaviour and as a result, this manifests discontinuous trajectories for the economic system. But for global, continuous processes (interest rates are only calculated at a certain time - the rates are continous), how does one account for zero-inflation processes?
We have anti-inflation in technology, due to the exponential growth of computational power at constant cost. A side-effect is the infinitesimal cost of second item e.g. software, music, video etc....
Time-delayed differential equations oscillate - and I hypothesize the huge diversity of modern time-constants (where some things are slow, others ultra fast) has reached beyond previous decades of thought in this area, due to the lack of new tools.
I know how to prove my molecules are behaving right - I do not think the econmists can say the same thing...
(Score: 3, Funny) by frojack on Monday May 09 2016, @06:54PM
I know how to prove my molecules are behaving right - I do not think the econmists can say the same thing...
But its just your unfamiliarity with the entire field of Economics that leads you to the false assumption that they care to make such a determination, - and more importantly - that economics operates with an assumption that there is such a thing as "behaving right".
You might just as well grouse that the fish don't pedal the bicycle in the most efficient way as complain that the Economist can't measure the "rightness" of individual behavior.
No, you are mistaken. I've always had this sig.
(Score: 2) by opinionated_science on Monday May 09 2016, @07:26PM
I modded you funny, because you're right! I know nothing formal about economics, but when a professor cannot defend themselves with anything other than hand waving, I feel my grasp of precision is perhaps moot....
There's always room for improvement, and I'm holding out for some computer models ;-)
(Score: 0) by Anonymous Coward on Monday May 09 2016, @01:23PM
His hand waving made her suspicious.