Ramanujan's easiest hard infinity monster (Mathologer Masterclass)

Video Statistics and Information

Video
Captions Word Cloud
Reddit Comments
Captions
Welcome to a Mathologer masterclass  video. We haven't had one of these   for a while :) This masterclass is devoted  to understanding a stunning mathematical   identity due to the genius mathematician  Srinivasa Ramanujan. Have a look at this:   1 divided by … and now it comes: 1+1  divided by 1+2 divided by 1+ 3 and then 4,   5, 6 all the way to infinity. A very curious  infinite fraction monster. Then there is also   this infinite SUM monster. And the pattern  for this second monster is obvious, right?   longer and longer products of consecutive odd  numbers in the denominators. Okay, now let’s   add the two monsters There, there. All this is  equal to, brace yourself :) …. Well, there’s pi,   because there’s always pi :) And there’s a factor  of e. Really ? Well, not quite finished yet.   Square root of pi times e over 2.  Hands up anyone who guessed that?   Nope, not a lot of hands. Seriously, what  a crazy result. How on Earth do those two   very different infinite monsters on the right  sum to this very simple product on the left,   tying together the numbers pi and e? Of course  Ramanujan was a genius at this kind of thing,   and our identity is just one of the thousands that  he discovered. Some people consider the identity   here to be Ramanujan’s most beautiful. Others  think of it as the “simplest” among Ramanujan’s   very hard results. Clearly “simple” is very  relative here. In 1914, the year that Ramanujan   left India for the UK., he challenged the  mathematical community to prove that his identity   is really true. There it is, as problem no. 541  in the Journal of the Indian Mathematical Society.   Today’s mission is for us to rise to Ramanujan's  challenge, and to puzzle out how these two monster   jigsaw pieces came together in the mind of  the mathematical genius. Along the way, we’ll   encounter some of mathematics’ most celebrated  results such as the Wallis product for pi.   The Gaussian integral And an infinite  expression that’s almost as crazy as   Ramanujan’s infamous deep connection between  -1/12 and the infinite sum 1+2+3+ and so on.   As usual lots to look forward to. But a warning.  To be able to understand everything that follows,   knowing some baby calculus would be really  helpful. No calculus? No problem :) Just   enjoy the mathematical fireworks courtesy of  one of the smartest persons who ever lived.   OK, time to begin. An infinite sum and an infinite  fraction, which one first? Definitely the sum,   right? What does that infinite fraction even  mean? So let’s start with Ramanujan’s sum. Some   of you will be familiar with a closely  related infinite sum, that one there.   So the denominator products are now consecutive  integers, not just the odd numbers. Okay,   so as a warmup, let’s make sense of this  second sum. Just relax and watch. Should   all look very natural to those of you  who’ve dealt with infinite sums before.   Let’s first turn this sum into a  power series in the variable x.   Make a copy. Let’s calculate the derivative of  this power series. Some baby calculus required   here. There, the derivative of x^4 is 4 x^3 Do  the same with all the other terms. simplify a bit.   Now the bit in the front That’s 1. Okay, lots of  coincidences. Can you see where this is going?   Align the top and the bottom. Okay, so the top  is a function in x, let’s call it y(x). Not a   terribly original name but it will do :) So  y on top and y’ its derivative at the bottom   The top is contained in the bottom Nice :) And  that means that we can replace the orange bit by   y. That’s a differential equation. To solve it, we  want a function for which, when you add one, you   get the derivative of the function. Are there any  simple functions that satisfy this differential   equation? I’m sure many of you will have  already guessed. Well, if the + 1 wasn’t there   our differential equation would say that our  function is equal to its derivative. Now,   which super famous function is the same as its  derivative? The exponential function, of course,   e to the power of x. Now, the exponential function  is not the only function that is equal to its   own derivative. 2 times e to the power of x is  also equal to its derivative. and so is 3 e to   the power of x. In fact, the functions that have  this special property are exactly the infinitely   many functions of the form constant times e  to the power of x. Now what if we throw the   1 back into our differential equation? That kind  of thing can create havoc with solving a DE, but   here it’s easy. We can compensate for that +1 just  by subtracting 1 from c times e to the power of x.   There in the orange, those are all the solutions  of our differential equation. But which of those   infinitely many functions is equal to our power  series? Well, that’s easy to figure out. Just   plug in x = 0. Then the right side just goes  away. And so we see that y at 0 is equal to 0.   Plug in 0 at the bottom Okay, y at 0 is 0. and e  to the power of 0 equals 1 and so c is equal to 1.   Quite a few of you will have seen this kind of  trick before. Now, finally, plug in x=1 and we   are done. This means that our warm-up infinite  sum amounts to e - 1. Pretty magical the first   time you see this. Okay, warmup’s done. Let’s  now try to make sense of Ramanujan’s infinite   sum in exactly the same way. There, all odd  numbers again, that’s Ramanujan’s sum. Extend   this sum into a power series, using just odd  powers of x. Can you see why that’s a natural   choice? Pretty natural, right? Well, let’s keep  going. Make a copy. Calculate the derivative.   The bit at the front that’s 1 again. Okay,  now how do things line up this time around?   Very good. But what about the powers? Slightly  off. But that’s easily fixed. Just pull out an x   at the bottom. As before, we’ll name the top power  series y(x), with the bottom its derivative, y’.   And, so, as in the warm-up, we can replace the  orange by y. The differential equation we get   here is pretty similar to the warmup one. The  only difference is the extra x on the right.   Okay so what are the solutions? What functions  will solve this differential equation?   Well, Marty just taught my first year students  here at Monash uni how to solve a differential   equation like this. As in the warm-up, there are  infinitely many solutions, and the value of y at   x = 0 pins down which of these solutions is equal  to the infinite series. Okay, so what’s the value   at 0. As before the right side wipes out and so  again y at 0 is equal to 0. Great. Unfortunately,   Marty’s gone on holidays, so he’s not around to  teach you how to solve this sort of differential   equation here. Anyway, there are general methods  that work, and again, you can sort of guess the   solution. If you want to give guessing a go, here  is a first Mathologer challenge for you: get rid   of the 1 and see if you can solve the resulting  simpler equation and let us know what you come up   with in the comments. But let’s get on with it,  and let me just show you what the solution is.   Whoa. Looks scary, doesn’t’ it? But it’s actually  not too bad. Don’t worry and just run with it for   the moment :) We’ll have a close look in a second,  but first, let’s finish our current mission,   to figure out Ramanujan’s sum. Okay, so our  differential equation work so far has given us   this identity. Remember what comes next? Yep,  to get Ramanujan’s sum we simply plug in x=1.   So, Ramanujan’s sum on the right is equal to that  weird integral thing on the left. Interesting :)   We can calculate the integral there, at least  numerically, and the whole left-hand side comes   to around 1.41. Aha! Very familiar, right? Maybe  the whole left-hand side is exactly the square   root of 2? And … nope. There is no aha. Sorry :)  The left-hand side is NOT exactly root 2. Too bad,   root 2 would have been a dream come true. But  still, we’re getting somewhere. Let’s look again.   e to the power of 1/2 that’s root e. So, a REAL  aha this time. Root e is one of the components of   Ramanujan’s identity. There root e is part of the  rooty thing on the left side of our identity. So,   we seem to be getting somewhere. What about the  integral? That will also look familiar to many of   you. Right? Normal distribution, bell curve,  important maths. What about the whole thing,   the integral? Well, that’s just the area  between 0 and 1 under the bell curve.   And that bell curve should definitely ring another  bell. Okay, okay cringy joke :) Bell curve ringing   a bell, sssss. The super famous Gaussian integral  that I mentioned at the beginning of this video,   remember? The total area of the bell, from -∞ to  ∞ is famously equal to the square root of 2 pi.   And half that is … :) Big AHA :)  Square root of pi over 2. Looking good,   right? Multiply the two roots. That’s EXACTLY  the left side of the identity we are chasing.   We're definitely on the right track. What’s  next? Well that yellow area splits into two bits,   the area between 0 to 1 which arose in the  Ramanujan sum, and the tail bit beyond 1. Well   that yellow area splits into two bits, the area  between 0 to 1 which arose in the Ramanujan sum,   and the tail bit beyond 1. Expressed in integrals  this splitting into two areas is just a sum.   Now pull root e inside the brackets. OK, so  remember, we’re chasing Ramanujan’s identity.   And we pretty much have it cornered. We’ve already  shown that the first green-yellow bit is exactly   Ramanujan’s infinite sum. So, what remains to show  is that the second green-yellow bit exactly equals   Ramanujan’s infinite fraction. Very, very nice and  ingenious. But now how can we show that the value   of the infinite fraction is really that second  green and yellow bit? Not obvious at all…. Hmm….   I’ve got an idea. Let’s switch to genius mode :)  Okay genius mode on. It’s that simple. Well, maybe   for Ramanujan… :) Throwing an x at the problem  worked really well for the infinite sum. So, let’s   try to go backwards, extend everything in front  of us to an x identity in a natural way and then   try to construct that infinite fraction starting  from that second green and yellow bit. Okay, split   up the left rooty side again. Now remember that  1 came from an x And that root e came from this   That’s a fairly natural thing to do, right?  And, most importantly, this new identity is   definitely still true for all x not just x=1.  Why? Well, because the two yellow integrals   will always add up to the yellow root pi over  2 on the left, no matter what x we choose.   Nifty :) Again, just to make sure that we are  all on the same page let me say this again:   At this point we know that the identity up  there is true for all x. Nothing to prove   anymore in this respect. Done and dusted :)  What remains to be done, in genius maths mode,   is to show that the function in the right box  can be written as an infinite fraction. AND   that this fraction becomes Ramanujan’s fractions  when we set x=1. Okay, so now let’s focus on the   fraction function. I’ll be lazy and call it y  again. Solving for y we get this difference.   What comes next going backwards? Well, what  about a simple differential equation for y?   To find the differential equation, we need the  derivative of y, and so let’s take derivatives   on both sides. Hard? Not at all. It’s really just  a matter of shovelling symbols around using baby   calculus. Piece … of ….cake. Here is what you  get. Pretty sure many of you will immediately   spot the gigantic coincidence. There, all of y  is contained in its derivative. and so we can   replace the orange bit by y. Looks very familiar,  doesn’t it. This new differential equation for the   fraction is very similar to the one for the  sum. Only difference is the minus instead of   a plus. Cool :) As usual, this differential  equation has infinitely many solutions. Now,   what do we usually do to pin down which of these  infinitely many solutions we are dealing with?   Well, by now you all know the drill:  We calculate y at 0 :) Okay so plug   in x = 0. Now what’s that on the right? Well, the  integral is from 0 to 0 and so the second term is   equal to 0. Now what’s that on the right? Well,  the integral is from 0 to 0 and so the second term   is equal to 0. And e to the power of 0 that’s  just 1. And so y at 0 equals root pi over 2.   That all works like a charm: and so the function  y is the special solution of our differential   equation that takes on the value root pi over 2  at 0 :) Final stretch. Let me now show you how   you can get that infinite fraction from this new  differential equation. Absolute magic, promise.   Okay, still in genius mode, to derive the monster  fraction, we start with the new differential   equation and first take derivatives on both sides.  On the left, the derivative of the derivative is   the second derivative. The derivative of -1 is 0.  And the derivative of x times y? Well, that’s just   the product rule from baby calculus, right? And we  get x times y’ plus 1 times x. Now repeat. Repeat   forever and ever after. Aha! Notice the  1, 2, 3 on the right? Ramanujan’s 1, 2,   3, fraction is just around the corner :)  Now fasten your mathematical seat belts.   Here comes some real mathematical  magic. Divide the first equation by y   And do the same for all the other equations. Have  a close look. Can you see what comes next? There,   there, and there. Yes, I can hear the yells for  me to substitute. And I’ll do that, but first   a little more aligning using algebra autopilot  before we engage the SUBSTITUTION autopilot :)   How amazing was that? And it looks like we’re  just about done. In fact, it looks like we   proved a lot more that what Ramanujan challenged  us to do. Right? We know this is an identity that   works for all x. THEN we know that the first  box is equal to this infinite power series.   and now we also know that the second box is  equal to this infinite fraction. Which lands   us with this identity. We can choose x however we  like, and we get Ramanujan’s challenge identity by   subbing x=1. Tada :) And we’ve made it. Impressed?  Relieved? I sure hope so :) However … Sadly,   there’s a however :( This more general identity  is definitely true and Ramanujan himself was very   much aware of some version of this identity.  But you regular Mathologerers will know that   infinity can easily fool us, and for what we’ve  done so far to really count as a proof, there are   still a few things that need to be checked. And  that’s what I warned you about at the beginning,   that we’d be entering the quicksand territory  of crazy things like 1+2+3+and so on “equals”   -1/12. So, we’ll finish with that, pointing  out the quicksand and how to avoid it.   Okay, so what’s the problem? To see the problem,  we have to recall how we came up with this   infinite fraction. Well, we began by noting  that y satisfies this differential equation.   and that y at 0 equals root pi over 2. Now, we  found ONE solution of the differential equation   in the form of our infinite fraction. However, we  did this SOLELY by manipulating the differential   equation, nothing else. In particular we did not  use the fact that y at 0 equals root pi over 2.   This tells you that there is more to be done,  right? After all, how can we be sure that the   solution we found is THE solution that also  takes the value root pi over 2 at 0? OK,   problem isolated, but now also easily sorted,  right? We’ll just plug x=0 into our fraction   and hopefully hammer the resulting expression  into root pi over 2. Okay, plugging in … now!   And now we need a very, very good hammer. How  can we see that this very curious infinite   fraction has the value root pi over 2? Well,  let’s see. To get a feel for what’s going on,   let’s make things finite and chop things  off at some point, say after the 7.   Now look at the bottom bit, the 5 and 6 and 7.  That’s the fraction 6/7th in the denominator,   with 5 in the numerator. So, we can flip and  multiply. And so the bit in the red box is this   Now, keep flipping until you  run out of things to flip :)   So, with fingers firmly crossed, the whole  infinite fraction should be equal to this   infinite product. All even numbers on top and  all odd numbers at the bottom. Does this remind   you of that other infinite product that I showed  you at the beginning of this video? This infinite   product here which is equal to … remember ? …  pi over 2 Super big AHA :) Again, that’s the   famous Wallis product. In the Wallis product all  the integers appear twice, 2 time 2, 3 times 3,   and so on. But, in our new infinite product each  integer occurs just once. What does that suggest?   Well, taking the square root of the Wallis  product should give us the value of our product,   right. And so our product should equal root pi  over 2. Exactly what we were hoping to find.   YESSSS :) Wallis’s product was the very good  hammer we needed and Life is good again :) That   was pretty crazy, wasn’t it? But here’s something  even crazier. It turns out that our new infinite   product is NOT equal to root pi over 2. What? Yes,  if you actually try to evaluate the product by   going 2 over 1 times 4 over 3 times 6 over 5 and  so on you’ll find that this infinite expression   explodes, … diverges to infinity. Another  Mathologer challenge for the keen among you:   try to pinpoint why the Wallis product does not  explode. Anyway, as you can see, we are in a   situation that is very reminiscent of Ramanujan’s  infamous 1+2+3+ “identity”. which really does not   make sense as an identity but which at the same  time captures a deep connection between 1+2+3+   etc. and -1/12, a connection that really works  like an identity under certain circumstances.   And the same turns out to be true for our infinite  product here. These impossible “identities” really   work in our scenario here, in that the infinite  function fraction, properly treated, really is the   correct solution to that differential equation,  and Ramanujan’s general x identity really is true.   Too deep and too crazy to go into here, but in the  end it all works out. Ramanujan’s identity really   delivers: mathematical genius, mathematical magic,  and real magic :) What more do you want? :) The   inspiration for this video was a 2020 blog post  by John Baez and the construction of the infinite   fraction was contributed by Leo Stein. Very nice  stuff. I’ve included links to the blog post plus   lots of other relevant articles and Mathologer  videos in the description of this video. Finally,   I note that x = 0 is a critical value for the  general formula up there. The formula only   works without invoking more mathematical magic  for x > 0. Also included in the description of   this video are articles that cross all the t s and  dot all the i s of this amazing formula. Anyway,   I hope you enjoyed this reconstruction of what  may have gone on in Ramanujan’s incredible mind   when he challenged the rest of the world to make  sense of his wonderful identity. Let me know in   the comments what worked and what didn't work for  you. And that’s it for today. Until next time.
Info
Channel: Mathologer
Views: 149,282
Rating: undefined out of 5
Keywords:
Id: 6iTdNmDHfV0
Channel Id: undefined
Length: 26min 45sec (1605 seconds)
Published: Sat Jun 24 2023
Related Videos
Note
Please note that this website is currently a work in progress! Lots of interesting data and statistics to come.