Tag Archive for 'Mathematics'

Three talks and many insights…

The last week or so has been quite eventful at TIFR. We have had some famous/important visitors who have given public lectures and talks.

CO2 Laser photo (From wikipedia)

The first visitor was Prof. Kumar Patel of UCLA, the Indian origin inventor of the $CO_2$ laser. He gave two talks. The first talk was for graduate students where he shared his insights and gave us some tips. The next talk was about about using lasers to detect explosives and chemical warfare agents. Both talks were interesting but I want to share some of the tips he gave us during his freewheeling, informal conversation with graduate students:

  • Be extremely ambitious and try to solve tough problems. Its no point trying to do old, solved problems a new way. Instead, go for fresh problems. Often, the greatest advances have come from scientists who solved problems that others would have considered impossible but out of inexperience the scientist never knew that! Experience can also prove to be a hindrance because it teaches you that you “can’t” do a problem in certain ways. Young people bring a certain irreverence to research that helps them make progress on so-called “impossible” problems.
  • What level of research should you pick? Should you pick a system that is simplified to its core or should you study a system in all its beautiful complexity? There is no correct answer. By stripping a system to its fundamental components and modeling its essential features you can learn a lot. But you can loose understanding of the collective behavior of the components. A beautiful example is the human brain…you must understand its basic component, the neuron and their collective behavior. A complete understanding of the human brain is not possible without understanding both. You cannot understand collective behavior of the neurons (e.g. thinking, dreaming, planning etc.) without understanding the basic components and vice-versa. Therefore what should happen is an integrative approach to science: Some people work from top down and others work from bottom up. Both will have something valuable to teach each other. (My idea: We see this in physics today. Astronomy tries to understand huge aggregates of particles in the form of galaxies, stars, planets etc. Particle physics tries to understand the fundamental particles themselves. Both disciplines feed off and fertilize each other. What aspect of science you choose to study i.e. the building blocks like cells, DNA, protons, quarks etc. or complex systems like the human brain, the weather, galaxies, the immune system etc. is a matter of personal taste. All of them are important, worthy pursuits)
  • I asked Prof. Patel whether science is for young people only (He invented the laser at a young age of 26). (I’m sort of an “older” student who has commenced his Ph. D. so I was personally interested in knowing whether I could still make an impact. It helped me that my question also was very applicable to him because he was in his sixties!) The answer he gave was fascinating: He said that the work he had done in the last six years was work he could never have done in the past. Its always a trade off he seemed to imply. At a young age you know a lot about a narrow area and you don’t know what is impossible. At an older age you know something about everything but are burdened by “this can’t be done” and so on. Also you just can’t keep up with the level of detail in research. (My idea: At a young age you should pursue a specific line of research to every level of detail possible. At an older age your ability to keep up with this level of detail may be lost but you can think about science which combines different areas. Here your experience helps you).

Lorenz Attractor

The next talk was by a French Prof. Étienne Ghys on the Butterfly Effect. The Butterfly effect is popular description of chaos: How a small event can have extremely large impacts on a system. “Does the flap of a butterfly’s wings in Brazil set off a tornado in Texas?” is a one of those questions that have entered the public imagination. I have my own version of the Butterfly effect (intended to be a joke of course)

[Sid Butterfly Conjecture 1] No matter how insignificant or poorly cited your research paper is, like the flap of a butterfly wing that causes a tornado in Texas, your paper will eventually have a earth shattering impact on the world of science :-)

This has another humorous implication:

[Sid Butterfly Conjecture 2] All research papers whether by Einstein or obscure researchers have the same impact eventually.

Incidentally I shared Sid Butterfly Conjecture 1&2 with the French Prof when I bumped into him. He seemed to love it!

Abel Prize Logo

The third talk that I want to mention was by S.R. Varadhan a famous Indian origin mathematician from ISI Calcutta who is now at Courant Institute, NYU. S.R. Varadhan was the winner of the Abel Prize (considered the Nobel Prize for Mathematics) in 2007. I have blogged about him here.

Hoping to learn from him, I asked him about the secret to his success. The answer was “you should have passion in what you do”. True, it was boring answer but its worth mentioning here because sometimes the secrets to success are quite simple. We shouldn’t expect rocket science answers to everything. Too often, the expectation of rocket science answers or tricks means that we don’t want to concentrate on the basics (which are obvious but tough to implement) like hard work, perseverance, passion, excellence, time management and so on.

It was inspirational to hear from two Indian origin scientists who have reached the heights of success in their fields. Can I apply their suggestions in my life? Only time will tell.

Online version of MATLAB/Mathematica/Maple/…

Sage screenshot

I’m very excited to talk about an open source mathematics system: SAGE.

SAGE aims to be an open source replacement for MATLAB/Mathematica/Maple. Whats amazing about Sage is the great functionality it gets by working nicely with already available open source math software (Maxima, Numpy etc). Its cute slogan “Building the car instead of reinventing the wheel” summarizes its software reuse philosophy. Because SAGE incorporates many different software projects, its quite complete (though it may never be as consistent or clean like a Mathematica or MATLAB). SAGE uses Python which possibly makes it the only computer algebra system that uses a mainstream computer programming language. The use of Python gives SAGE tremendous flexibility and power.

One of SAGE’s most amazing features…which is actually the main point of the blog…is that you can use it online!! This is really cool because you can do this from a browser anywhere on the Internet. In the future, if you are stuck on a computer which does not have MATLAB/Mathematica, despair not for you can use SAGE.

The SAGE online interpreter is available here. The style of SAGE is a bit like Mathematica. You enter an expression into Notebooks and type Shift+Enter to evaluate it. You can do all kinds of nifty things like collaborate with others and publish your notebook on the web.

___

Nice introductory video on SAGE. Guaranteed to get you all excited…

SAGE according to Wikipedia

Another screenshot

SAGE screenshot 2

SAGE Logo

SAGE Logo

Technologies in SAGE

SAGE Technologies

Navier Stokes Equations – An Introduction

navier_stokes.gif
Airflow is simulated over and past the wing of a high performance aircraft that is using vectored thrust while descending to a few feet above the ground (in ground effect). See here

The Navier-Stokes Equation is a partial differential equation that models fluid flow. Fluids include liquids and gases. Using Navier Stokes we can understand diverse phenomena like airflow over aeroplane wings or rocket bodies, the flow of liquids through pipes and the flow of plasma in stars (magnetohydrodynamics). There are many computer programs available that will divide a fluid flow problem into discrete, finite sized pieces and solve the problem for you (the so-called finite element analysis method). NASA, major defense contractors and car manufacturers all use these programs to solve the sort of problems listed above.

Navier Stocks Equation - Incompressible Liquid with Constant Viscosity (Newtonian Liquid)

Above: Navier Stocks Equation – Incompressible Liquid with Constant Viscosity (Newtonian Liquid)

It may come as a surprise that even though Navier-Stokes in used so extensively in physics, engineering and industry we know little about the theoretical underpinnings of these equations. Anybody who can answer certain fundamental questions about the Navier-Stokes equations can win a \$1 million award from the Clay Mathematics foundation.

The lack of adequate knowledge about the Navier-Stokes equations represent one of the greatest unsolved problems in 21st century mathematics. The announcement of the \$1 million award certainly attracted my attention but I found Navier-Stokes equations to be so interesting that I taught myself some basics. I would like to share some useful resources:

A good introduction to Navier-Stokes equations can be found on the following links:

(1) Navier Stokes Derivation

(2) Navier Sokes Equations

(3) A infinitesimal element approach to deriving Navier Stokes

(4) A slightly bizarre but fun link

But the pièce de résistance has got to be the video* of the problem description by the Clay Mathematics Institute. Its an elementary description of the problem that anyone with an understanding of vector calculus can comprehend.

Tip: Read links (1) & (2) of wikipedia before listening to the lecture. You will gain more out of it. In the lecture Cafarelli uses non-standard notation to denote derivatives. For example, D_x v means \frac{\partial v}{\partial x}.

__

(*select Navier Stokes existence and smoothness by Luis Caffarelli, University of Texas)

Math Joke

Picked up from the internet…

What did i say to \pi?

“Be rational!”

\pi replied…

“Get real!”

Image

MATLAB tip: Avoid loops and use surf command

Frequently we will get a function f(x,y) that we will need to plot. The naive way will be to generate an array for x and y and then iterate through all combinations and then use the plot3 function. Here is a sample:

  1. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  2. % Assume function f(x,y)=-(x^2-2)^2 - (x^2-exp(y))^2
  3. % Crude Method
  4. x = -5:.1:5;
  5. y = 1:.1:4;
  6. points=[];
  7.  
  8. i=1;
  9. for xp = x
  10. for yp= y
  11. f = -(xp^2-2)^2 - (xp^2-exp(yp))^2;
  12. points(i,1) = xp;
  13. points(i,2) = yp;
  14. points(i,3) = f;
  15. i=i+1;
  16. end
  17. end
  18.  
  19. figure;hold on;grid on;
  20. xlabel('x');ylabel('y');zlabel('f');
  21.  
  22. plot3(points(:,1),points(:,2),points(:,3),'.');
  23.  
  24. hold off;

But this code is inefficient. Looping is to be avoided in MATLAB because operations on matrices are faster. Moreover, the plot3 function does not give shading. We get the following as the output:

Graph Using plot3

Avoiding loops, we can implement the above code in a different way. We avoid loops and use the surf command.

  1. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  2. % Assume function f(x,y)=-(x^2-2)^2 - (x^2-exp(y))^2
  3. % Better Method
  4. x = -5:.2:5;
  5. y = 1:.2:4;
  6.  
  7. x_2d = ones(length(y),length(x))*diag(x);
  8. y_2d = diag(y)*ones(length(y),length(x));
  9.  
  10. func = -(x_2d.^2-2).^2 - (x_2d.^2-exp(y_2d)).^2;
  11.  
  12. figure;hold on;grid on;
  13.  
  14. xlabel('x');ylabel('y');zlabel('f');
  15. surf(x,y,func);
  16.  
  17. hold off;

We get the following output

Graph using surf and no loops

The trick in the above code involves creating the x_2d and y_2d array.

References:
(1) plot3
(2) surf

Quick Introduction to Manifolds

Charts and Atlases

My M.Sc. Project at IIT Madras deal with Manifolds and some numerical computations on them.

The initial part of my M.Sc. project introduced Manifolds in general. I’ve put that up on the web. So if you are interested in a quick and dirty introduction to Manifolds, here it is! There are lots of diagrams and figures to make the concepts easy. Its a fun and fast tutorial that can be completed in a few hours.

Here is the Preface

This HTML document aims to introduce Manifolds. It has been derived (mainly verbatim) from parts of my M.Sc. Physics final project. I used Latex2HTML to convert my latex to the HTML you see here.

Manifolds are fundamental structures in Differential Geometry. The study of Manifolds is useful in various branches of Theoretical Physics, especially High Energy Physics and General Relativity. For instance, Einstein’s theory of General Relativity conjectures that space-time forms a 4 dimensional pseudo-Riemannian Manifold. Superstring Theory explains the compactification of extra dimensions by using Calabi-Yau Manifolds.

Manifolds are abstract mathematical spaces that look locally like \small{\mathbb {R}^n} but may have a more complicated large scale structure. The surface of Earth is a simple example: At small distances it looks like the Euclidean \small{\mathbb {R}^2} but from far away it is \small{\mathbb {S}^2}, the two dimensional surface of a sphere. The behaviour at the small scale and large scale can be totally different. For instance, in \small{\mathbb {R}^2} parallel lines never meet while all lines eventually meet in \small{\mathbb {S}^2}. Because all Manifolds are locally like Euclidean Space we can develop common mathematical techniques to study extremely different kinds of spaces.

We can define increasingly complicated structures on Manifolds so that we may do Calculus on them or define concepts of distance and angles on them. We may also want to study Manifolds in terms of complex variables and perform Complex Calculus on them. In this report we look at the whole hierarchy of Manifolds. We start from Simple Manifolds and progress to Differentiable Manifolds, Riemannian Manifolds and lastly Complex Manifolds. Within Complex Manifolds we study Hermetian Manifolds and Kahler Manifolds. Orbifolds, another special kind of Manifold, are also introduced. All the related mathematics and concepts such as Vector Fields, Tangent Spaces, Metrics, Curvature, Parallel Transport and Connection are explained.

Still interested? Here is the HTML document

Indian wins big math prize

prisvinner_2007_lite.jpg

In March 2007 an Indian won a big award in Mathematics. I had written about it at IIT Madras (for a mailing list). I reproduce the writeup here:

The 2007 Abel Prize has been awarded to S.R. Srinvasa Varadhan (Faculty NYU) (Ph. D. ISI Calcutta 1963) for his contributions to probability theory.

The Abel prize is worth approx USD 997,000 and is presented by the King of Norway. The winner is selected by the Norwegian Academy of Science and Letters. Some people say that the Abel Prize is the Nobel equivalent for Mathematics (Though this can be disputed by those who would give that place to the Fields Medal which has a longer history. Interestingly the fields medal is only awarded to mathematicians under 40. This, I believe makes it
a “flawed” prize. Also, the fields medal would only make you richer by US $13,000).

The Abel prize is named after a Norwegian mathematician Niels Henrik Abel who made many impressive contributions before he died at the age of 26(!).
(When something is commutative i.e. a*b = b*a we also call it Abelian in his honour). What’s interesting to me is that Abel rhymes with Nobel :-) Perhaps that bodes well for the prize in the public’s imagination.

S.R. Srinivas did his Ph. D. in ISI Calcutta (1963) and then went for his postdoc at the Courant Institute of Mathematical Sciences at NYU. Incidently he is from Chennai and did his B.Sc. from Presidency (1959).

I have found Tamil translations on the website of the prize. Go to
here for a tamil translation.

Lets go for the big prizes!! We Indians can do it (perhaps not in
cricket??) !

For those who are really interested

The authoritative source
http://www.abelprisen.no/en/

More Info:
http://en.wikipedia.org/wiki/Abel_Prize

Some news articles:
International Herald Tribune

The Hindu
Minor Quibble: The Hindu incorrectly lists the amount as US $850,000. Its
about US $975,000