I'm a middle-aged guy who's been studying this stuff at a very amateur level for the last couple of years. It's so much fun and feels like learning the secrets of the universe. I just wish I were young and could study this in grad school... if there's anything left to study. :)
Mathematics is a young man's game because older men think mathematics is a young man's game. Mark my words: keep at your "amateur" study, otherwise one day you will find yourself saying, _"I'm an old man...I just wish I were middle-aged..."_ 🤓
I am also interested in the relation between number theory and complex analysis.. Can u please share ur telegram number Or email so that we can discuss about it??
Don't u DARE give up friend! I went back to school as a challenge at 36 & though I'm 52 now, I have 2 Math /Applied Stats Degrees & hoping to at least get a Master's once the $$ can be balanced out. It was demanding & I did FAIL /repaeat some higher Analysis courses, but it was all worth it...Math is simply too BEAUTIFUL for me to not keep trying...LOL
This video explains one particular theme very well in which Ramanujan's 1916 paper called " On certain Arithmetical functions " had an impact on the eventual proof of Fermat's Last Theorem, through the works of Hecke , Mordell , Eichler and Shimura and later by Deligne who proved one of Ramanujan's Conjecture by associating to that Delta Modular form of Ramanujan a Geometric object called a Motive ( à la Grothendieck ). I'd like to mention another theme from that 1916 paper which is even more directly related to the Wiles' proof of FLT. On one hand where Ramanujan conjectured some direct properties of his Delta Modular form - like the fact that its coefficients are multiplicative, he further proved some congruences related to the coefficients of Delta which were extremely bizzare and completely unexpected at first. One such congruence which occurs in that paper was - ' tau(p) is congruent to 1+p¹¹ modulo prime 691 ' , here tau(p) is the pth coefficient in the expansion of delta. It was Jean Pierre Serre who realized that there has to be some reason behind these congruences and their existence. To explain these, Serre discovered a huge set of ideas - he developed the notion of what's called p-adic and mod-p modular forms, related them to ' mod-p Galois Representations ' ( an extremely important tool in modern Number Theory ) , gave a new definition of what's called p-adic Zeta function which is itself related to an old approach of Ernst Kummer to prove Fermat's Last Theorem whenever exponent in the Fermat's equation is a ' regular prime ' and lastly , while he was trying to explain Ramanujan's congruences , Serre formulated what came to be known as " Serre's Modularity Conjecture " , he further deduced Fermat's Last Theorem directly from his Conjectures without having to take the middle step of using Shimura-Taniyama Conjecture and later it turned out if we actually do wish to take that middle step then a small part of Serre's original Modularity conjecture would suffice to prove the implication :- Shimura Taniyama => FLT , Serre called this small part " Epsilon Conjecture " and that's exactly what Ken Ribet proved thereby paving a way from Shimura-Taniyama to FLT. Infact it doesn't end here yet , both - A sophisticated version of the theory of p-adic modular forms as well as a proved special case of Serre's conjecture, which Serre developed to explain Ramanujan's congruences was used by Andrew Wiles himself in his ' Modularity Lifting Criterion " which was the most important step in the proof of Shimura-Taniyama Conjecture. So overall , the influence of Ramanujan on the proof of Fermat's Last Theorem is much more than we think it is. Edit: By the way the full " Serre's Modularity Conjecture " is now a theorem of Chandrashekhar Khare.
@Prakhar Pratyush do you know who discovered that the coeff of the modular form expansion are related to the number of solutions mod p of an elliptic curve, and how on Earth he or she came to that discovery?
@@zy9662 There's actually a long tradition of people trying to find correspondence like these which are often called " Reciprocity Laws " starting from the work of Gauss on Quadratic Reciprocity to the work of Emil Artin on his Artin Reciprocity law which gives a similar correspondence for Polynomial equations in one variable whose Galois Group is abelian, so it's not something out of nowhere. The Reciprocity laws related to Elliptic Curves were a next step in this tradition first started by i guess Martin Eichler and Goro Shimura as explained in the video itself. Given a specific sort of Modular form there's a way to attach an Elliptic Curve to it and it can always be done but the converse was formulated first by Yukata Taniyama and Goro Shimura and then refined by Andre Weil.
@@pursuingstacks yes, but the reciprocities of Gauss until Artin are more about mapping of groups, but identities between integer sequences are a different breath altogether. Modular forms had been around in number theory way before Eichler paper and I was wondering if it was Eichler who found the correspondance or maybe Hasse.
This is the first time I've actually had the connection between elliptic curves and modular forms explained. I had always wondered about why elliptic curves were used in encryption, and learning of this connection explains why. I had understood that RSA used large primes, and that this was connected to a modular form. So this video helped me see the connection between RSA and ECC. Fascinating. I absolutely love videos that are capable of giving me these sort of insights. I do read a lot of Wikipedia articles and other sites when looking into math, but just reading words and starting at an equation simply cannot replicate the sort of insights I make during the course of a video such as this.
I'm not an expert in this area, but my understanding is that elliptic curves are used in cryptography because they have a group operation x.y -> z which is easy to compute in one direction, but not in the other. The connection between elliptic curves and modular forms is not relevant to the use of elliptic curves in cryptography. (Wikipedia link: en.wikipedia.org/wiki/One-way_function#:~:text=An%20elliptic%20curve,compute%20k.)
As someone currently studying arithmetic geometry and proof of Fermat's last theorem, this is simply fantastic. You really captured some of the special meaning of the Eichler-Shimura theorem and modularity simply, and in just 10 minutes no less! I'm astonished!
Thank you! Another gorgeous video. Your content is always really well done, and strikes a really, really impressive balance between parsimony and depth of insight. Please keep up the great work! Looking forward to the next one!
One of the most beautiful and aesthetically pleasing videos in have ever seen. I knew the maths, I knew the journey, but your video, both concisely, logically and especially visually was an absolute delight.
My master's thesis used a lot of modular forms. This was a really nice historical perspective and transitioned so cleanly from the definitions to their impact
1 year ago yt recommended me the video about derivatives, it's still hard for me but it sparked my interest in maths. Great job on the videos. I hope more people will see them in the future.
This has hot to be one of the all time best math videos I have ever seen.This is the first time I have been able to understand the connection between modular forms and elliptic curves
Yes, I knew that RSA was connected to modular forms, and that elliptic curves are commonly used in cryptography, but I wasn't aware that they are both connected through modular forms.
I like how you show the genius mathematical intuition of these peoples (even if these ideas took them years to conceptualise). Thank you for your videos!
The entirely unintuitive triumph at 6:12 literally made me shout in disbelief. It’s astounding that this isn’t an elaborate gag. Thanks for the great warm-up video.
At this point of time, I think I can safely state that this channel is (at least in my opinion) one of the best mathematics related channels on RUclips. Btw awesome video!
Very simplified, but accurate. I hope these videos will stimulate some people to take up the so called hard maths. Very rewarding. Looking forward to more..
I've always wanted a video that gave a little insight into modular forms that didn't immediately soar above my head. They always come up in all sorts of places and nobody ever even tries to give the faintest intutition... Thank you.
I'm actually doing my Ph.D. on a somewhat related topic (transfer operator techniques on hyperbolic surfaces, which can be linked to automorphic forms of which modular forms are a subset. There they appear as building blocks for eigenfunctions of the Laplacian). My impression as to why nobody seems to bothers to give an intuition about modular forms is that nobody really has one. People study them because they are quite peculiar things and seem to hint on profound hidden meaning whenever given the slightest of chances to do so. But apart from that, as far as I know, they remain quite mysterious things. They're like the Dark Matter of math.
@@lonestarr1490 They certainly are mysterious! The closest intuition I've found to modular forms being "natural" objects is that they are sections of line bundles on modular curves, and modular curves are moduli spaces of elliptic curves (which are definitely natural objects!). The way we study spaces in algebraic geometry is by understanding the functions that live on them, i.e. sections of line bundles (or vector bundles, more general sheaves etc.) so in this sense modular forms arise "naturally". Why this in turn has so many applications to complex analysis, hyperbolic geometry, sphere packings, quantum groups, etc. I have no idea, and I have to agree with your perspective that they're just the "Dark Matter of math"
Your integration(pun intended) of history, small facts and you not fearing of losing viewers because of big maths is really refreshing. We need ofc those dumbed down channels but we also need this, this is like a proper documentary made by BBC that shows the history of a place, showing it's founders and stuff or an Animal planet show etc. Such great job
Your channel won’t stop getting better. No other channel makes videos like yours, and any one that tried would be facing an uphill battle to compete with you.
There is actually another connection between a lattice and number theory. If you define a linear, recursive congruence of the form r_{i + 1} = (a*r_i + c) mod M, where a, c and M are constants, then plot successive elements of the sequence as points on a plane (x, y) = (r_i, r_{i +1}) you will see a lattice. If you choose c=1, a=4, M=9, and a starting value of r_0 = 3 you get the sequence 3, 4, 8, 6, 7, 2, 0, 1, 5, 3... and so on repeating. You then will have to plot (3, 4), (4, 8), (8, 6)... and so on. This example is taken from Computational Physics Problem solving with Python by Rubin H. Landau Manuel J. Páez , and Cristian C. Bordeianu, chapter 4.2.1. It is presented as the linear congruent method for finding pseudo-random numbers. In the book it is given as a bad example of one, considering that the series repeats quite early. I do not know how you would relate this further with elliptic curves, nor if it has been done already.
This is a great video. I believe a lot of patters and corollaries we see in mathematical disciplines has to do with the nature of computation. Any multiplication or division operation in a non binary numbering system generates a symmetrical form of operations and ultimately creates prime numbers. It will be cool when people start to crack some of the toughest problems in the fields. It's good to see that curiosity isn't dead!
I'm curious where the 24th power in the first equation comes from. I know that specific equation only serves as background to help introduce the problem, but whenever I see a specific equation with seemingly random constants in them I always wonder where the constant came from.
It's a very important 24 :). The simplest definition of the modular discriminant function (the delta) is delta = g2^3 - 27g3^2, where g2 and g3 are essentially the coefficients of the elliptic curve attached to a modular form (or lattice). Then, g2 and g3 are essentially Eisenstein series; g2 is a sum of -4th powers, and g3 is a sum of -6th powers. So delta ends up being something like a sum of 12th powers. Those are 12th powers of tau, a variable in the upper-half plane; moving tau to q (the "nome") in the unit circle throws in another power of 2. Thus, an overall power of 24. (There's much more, like that delta is the first nontrivial *cusp* modular form, and appears with weight 12, which then becomes 24 after converting tau to the nome. And I think the 24 can also be traced back to the special lattices formed by {1,i}, which gives a factor of 4, and {1, e^{2pi/3}}, which gives a factor of 6. Then 4*6 = 24. I think 4 and 6 are also the nontrivial conductors of imaginary quadratic fields, or something like that.)
@@pinkalgebra Thank you for the easy to understand explanation! I was also wondering why in the world 24 of all numbers was used, but this makes a lot of sense.
@@pinkalgebra well my last comment got deleted for linking to, of all things, a stackexchange post about numerically solving this stuff in Java... A lot of that initially went over my head, and I had to do a ton of digging through Wikipedia and other sites until I managed to piece together a semblance of understanding, but I'm beginning to understand what you're saying. I really appreciate the reply, as it has led me to quite a few interesting things to read, and has definitely helped point me towards a better poverty understanding of the topic. I mentioned in another comment here I'm self taught beyond pre-calc and the power rule for solving derivatives (and I have a tendency to not follow the standard order school trends to teach things, often diving into the deep end and trying to piece together an understanding of things). So while I still don't fully "get" the math, I can see where the numbers come from a bit better.
Very nice historically grounded exposition. RUclips puzzles me given that someone who is conversant with the fine details of many theorems mentioned is likely watching and yet there is in general no middle ground between light popularization and heavy research talks on this platform. Mathematicians are shy creatures, perhaps this furnishes an explanation, not to mention that RUclips is somewhat of an ad riddled commodification machine that seems almost designed to repel serious mathematical discourse. Perhaps we can aim for a sort of digital common room in the distant future. Science shouldn't be confined to the limits of earshot.
Great video indeed. Just watched a Terence Tao video few hours ago. I am lacking complex geometry and various maps knowledge. Elliptic curves occur in other number theory domains, for example congruent numbers, and even theta-congruent numbers, where the elliptic curve has a specific form. I am working on mathematical billiards and they occur there too.
This was a great video. On a side note which I think is super cool, I saw the picture of Conrad at the end and could have sworn it was my old Calculus 2 professor (Keith Conrad). Turns out they’re brothers, crazy. They’re nearly identical
Great explanation! I'm a bit confused at 5:43 though. For p = 2 how could there be 5 solutions if there are only 2^2 = 4 possible ordered pairs mod 2? Aren't the only solutions (0,0) and (1,0) since the right-hand side is always even?
I think the right elliptic curve is y^2+y =x^3-x (you may modified to make it the normal form) and the number of solutions should be modified to be +1(th infinity point is included)
It could be that it's a modular form of level divisible by 11. The theorem excludes a finite set of primes, namely those dividing the "level" and the auxiliary prime of your Galois representation. However, the pattern holds for all remaining primes.
So, a few questions that immediately spring to mind: Of course for any elliptic curve you can just make a list of all solutions mod all the p's and then write an f(q) with those coefficients, so... 1) what has to be done to prove that this is modular (like, is being modular any more special than looking like what you showed) 2) What has to be done to show that f(q), integrated over all the 'special arcs', gives a lattice which can be fitted periodically by a set of functions (x and y) satisfying the original elliptic curve ( y = f(x)) that it all came from? In other words, are these things comparably difficult to prove as all the other difficult things that you mentioned people were trying to prove here, or have I missed some 'obvious' logic?
Does multiplicativity only hold for coprime exponents? What's the meaning of the prime power coefficients? (-2q²)(-2q²) ≠ +2q⁴ (-q³)(-q³) ≠ -2q⁹ EDIT: probably also related to the number of solutions in a ring. That would also explain the multiplicativity, comes from ring factorization. But which ring? Cyclic? Galois field?
I may have to check your videos out. I've been wondering what the connection to the torus is, so hopefully your second video will be able to provide that insight.
@@Bobbias there's some wild stuff going on there. For FLT, I'll be doing a series of lectures much like for the Congruent Number Problem, but then I'll also turn everything into a much longer series of refined, shorter videos (10-20 min each) with a little higher production quality correcting any mistakes from the lectures.
@@garethma7734 You are correct, but additionally the equation is singular when reduced mod 2, 3, or 11, so there's some special-case handling needed for those primes.
Quick question: At 5:11, Aleph 0 states that a common question, when given an equation, is "how many solutions are there mod p," where p is a prime number. However, at 5:59, when Aleph 0 goes on to list p and the number of solutions to mod p, they do not include 11 as one of the p values. Was this just a simple mistake or does the number 11 not satisfy the condition that 1 + p- (# of solutions mod p) is represented in the coefficient of the modular form?
That's a great observation given how often it's overlooked in various expositions , which by the way makes sense given the technical nature of the question. If you read the description of this video at some point aleph0 mentions how those lattice points are obtained by paths on that unit disc which are sort of invariant under the action of something called Gamma_0(11). Notice how 11 shows up here. This number 11 associated to a Modular form is called Level of a modular form. It appears in the very definition. Moreover when we match our modular form with the elliptic curve what we are really looking for is to have an analogous notion of level on the side of elliptic curve as well so that they both the values of levels match. In the world of Elliptic Curves that notion already existed and is called " Conductor ". It measures the reduction type of your elliptic curve what that means is , for all prime numbers p dividing the conductor , the equation you get after reducing your Elliptic Curve mod p will no longer be non-singular and what that means technically is that the tools you have to measure number of primes modulo p on a curve no longer work for that non-singular curve. In our case since Level of modular form is 11 , the conductor of corresponding elliptic curve would be 11 as well , and the prime dividing this conductor is 11 itself , the so called ' bad prime ' and so reducing your elliptic curve mod 11 would give you a non singular curve and so you would want to avoid it.
P=11 actually is the starting solution node for FLT. The cycle repeats at P=10011, as on a torus . The solution using a secret coefficient node is e^(pi+8), such that the distance between exponents is 2807/1579=1.7777074, starting from p=11
For no particula reason? When harvest starships will come back to Earth and found Ramanujan's work and decide to save our species, you'll get the reason.
Great video, as usual! I have been fascinated with modular forms since I started reading about them in popular summaries of FLT and Wiles' proof. What are the steps towards getting deeply acquainted with them mathematically? I imagine an undergraduate course in Complex Analysis would be the starting place, wouldn't it?
Some complex analysis is required, but not that much. A fair amount of number theory is required. I suggest checking out Tom Apostol's textbook on modular forms. I believe he gives you the pieces of number theory and complex analysis that you need, and you can freely skip over the more technical details of stuff like contour integration which is required for some proofs. The main skill required may be being good at manipulating infinite series and infinite products. Good luck and have fun!
As someone mostly self taught beyond pre-calc, just seeing x = sin theta, y = cos theta, and x^2 + y^2 = 1 helped me see the connection between sin/cos and the unit circle in another way of never noticed.
Such a great topic! I've been just thinking about how you can tell the GCD of two numbers by looking at where two sine waves with the period of those two numbers meet at 0!
That has got to be one of the least satisfying ends to an aleph null video ever. Where can I find more? Also, if I may ask a few questions, what is your specific area of study? What schools did you attend? And where do you find your articles and sources for your videos? Also how did you get so good at explaining these harder or more advanced topics?
There are numerous textbooks on modular forms and elliptic curves. I started with Tom Apostol's number theory books, which has modular forms as a subject, but doesn't really do modular forms. There's a book by Lawrence Washington on elliptic curves that I'm finishing up now, which is probably my favorite intro to elliptic curves, and does only a little bit on modular forms (though a lot on the Weierstrass p-function). So those are some places to find more. You can also just read the Wikipedia and/or Mathworld articles on the things mentioned. Have fun!
Nice video, given I know nothing about this, I could follow it pretty well. I just need to clarify, what exactly is a modular form? It wasn't explained so clearly.
Clear and concise! I'm a new subscriber :) Just curious, what is the 'so what?' of this finding? I'm sure it's a broad impact connecting modular forms to elliptic curves, and it's connection to Fermats last theorum, but what would be a prominent example of the use of this discovery?
Are there any relations or info regarding the intermediate coefficients? For example at 0:43 you say -24*252 = -6048. Then soon after, *skipping* -1472 you say -24*4830 = -115920. And that's b/c the place/index of the coefficients, ordered, need to be coprime. So what about other coefficients?
The paper "On the modularity of elliptic curves over Q: wild 3-adic exercises" is about 100 pages. It will make good sense when it can be written easily over a few pages.
Modular forms and elliptic curves Infinite phi revolving around infinite parallels Fractals of infinite realities each cascading, gliding In an infinite wheel Tell me the true nature of my reality! These are the verses that made me want to become a mathematician.
Great video! But I still don't understand how you connect the delta(q) (e.g. delta(q) = q[(1-q)(1-q^2)(1-q^3)...]^24), the resulting lattice to elliptic curve equation then the function f(q) = q - 2q^2 - q^3 +2q^4+ q^5+2q^6-2q^7-2q^9-2q^10+q^11+4q^13....). Where does the last f(q) polynomial expansion come from relative to the first? I get that the solutions (1+p-) (# of solutions mod p) reveal the same coefficients as this equation, but don't grok where it comes from as one steps through your video. Seems to pop out of nowhere? (e.g. the f(q))
I'm a middle-aged guy who's been studying this stuff at a very amateur level for the last couple of years. It's so much fun and feels like learning the secrets of the universe. I just wish I were young and could study this in grad school... if there's anything left to study. :)
Number theory (and complex analysis, by extension) has a LOT of unproven conjectures and open problems, there is definitely more to study!
Mathematics is a young man's game because older men think mathematics is a young man's game.
Mark my words: keep at your "amateur" study, otherwise one day you will find yourself saying, _"I'm an old man...I just wish I were middle-aged..."_ 🤓
There is a 75 year old man in my grad school. You’re definitely not too old.
I am also interested in the relation between number theory and complex analysis.. Can u please share ur telegram number Or email so that we can discuss about it??
Don't u DARE give up friend! I went back to school as a challenge at 36 & though I'm 52 now, I have 2 Math /Applied Stats Degrees & hoping to at least get a Master's once the $$ can be balanced out. It was demanding & I did FAIL /repaeat some higher Analysis courses, but it was all worth it...Math is simply too BEAUTIFUL for me to not keep trying...LOL
It's hard to imagine that Ramanujan just stumbled upon modular forms. There was profundity in everything he touched, even when he didn't realize it.
He truly was a prodigy. Imagine what he might have discovered if he didn't die so young and senselessly.
@@Red-Brick-Dream an anomaly of nature is what I can best describe him from my baseless reminiscence of his works through books
Maybe it was his hypothalamus?
Ramanujan’s intuition for number theory was astounding and terrifying.
i feel like mine is just as good. just the low hanging fruit has all been plucked.
@@sharpnova2this is not low hanging fruit.Most math after the last 100 years isn't. Especially for its time.
@@sharpnova2 make sure to check that self-diagnosis dillema of humans
@@sharpnova2 💀
@@sharpnova2😭😭😭😭
This video explains one particular theme very well in which Ramanujan's 1916 paper called " On certain Arithmetical functions " had an impact on the eventual proof of Fermat's Last Theorem, through the works of Hecke , Mordell , Eichler and Shimura and later by Deligne who proved one of Ramanujan's Conjecture by associating to that Delta Modular form of Ramanujan a Geometric object called a Motive ( à la Grothendieck ).
I'd like to mention another theme from that 1916 paper which is even more directly related to the Wiles' proof of FLT.
On one hand where Ramanujan conjectured some direct properties of his Delta Modular form - like the fact that its coefficients are multiplicative, he further proved some congruences related to the coefficients of Delta which were extremely bizzare and completely unexpected at first. One such congruence which occurs in that paper was - ' tau(p) is congruent to 1+p¹¹ modulo prime 691 ' , here tau(p) is the pth coefficient in the expansion of delta.
It was Jean Pierre Serre who realized that there has to be some reason behind these congruences and their existence. To explain these, Serre discovered a huge set of ideas - he developed the notion of what's called p-adic and mod-p modular forms, related them to ' mod-p Galois Representations ' ( an extremely important tool in modern Number Theory ) , gave a new definition of what's called p-adic Zeta function which is itself related to an old approach of Ernst Kummer to prove Fermat's Last Theorem whenever exponent in the Fermat's equation is a ' regular prime ' and lastly , while he was trying to explain Ramanujan's congruences , Serre formulated what came to be known as " Serre's Modularity Conjecture " , he further deduced Fermat's Last Theorem directly from his Conjectures without having to take the middle step of using Shimura-Taniyama Conjecture and later it turned out if we actually do wish to take that middle step then a small part of Serre's original Modularity conjecture would suffice to prove the implication :- Shimura Taniyama => FLT , Serre called this small part " Epsilon Conjecture " and that's exactly what Ken Ribet proved thereby paving a way from Shimura-Taniyama to FLT.
Infact it doesn't end here yet , both - A sophisticated version of the theory of p-adic modular forms as well as a proved special case of Serre's conjecture, which Serre developed to explain Ramanujan's congruences was used by Andrew Wiles himself in his ' Modularity Lifting Criterion " which was the most important step in the proof of Shimura-Taniyama Conjecture.
So overall , the influence of Ramanujan on the proof of Fermat's Last Theorem is much more than we think it is.
Edit: By the way the full " Serre's Modularity Conjecture " is now a theorem of Chandrashekhar Khare.
Man, I love reading your comment even though I couldn't make out the head and tail of what you're commenting.
Thank you for this comment. It's very insightful and adds to this excellent video.
@Prakhar Pratyush do you know who discovered that the coeff of the modular form expansion are related to the number of solutions mod p of an elliptic curve, and how on Earth he or she came to that discovery?
@@zy9662 There's actually a long tradition of people trying to find correspondence like these which are often called " Reciprocity Laws " starting from the work of Gauss on Quadratic Reciprocity to the work of Emil Artin on his Artin Reciprocity law which gives a similar correspondence for Polynomial equations in one variable whose Galois Group is abelian, so it's not something out of nowhere.
The Reciprocity laws related to Elliptic Curves were a next step in this tradition first started by i guess Martin Eichler and Goro Shimura as explained in the video itself.
Given a specific sort of Modular form there's a way to attach an Elliptic Curve to it and it can always be done but the converse was formulated first by Yukata Taniyama and Goro Shimura and then refined by Andre Weil.
@@pursuingstacks yes, but the reciprocities of Gauss until Artin are more about mapping of groups, but identities between integer sequences are a different breath altogether. Modular forms had been around in number theory way before Eichler paper and I was wondering if it was Eichler who found the correspondance or maybe Hasse.
This is the first time I've actually had the connection between elliptic curves and modular forms explained. I had always wondered about why elliptic curves were used in encryption, and learning of this connection explains why. I had understood that RSA used large primes, and that this was connected to a modular form. So this video helped me see the connection between RSA and ECC. Fascinating.
I absolutely love videos that are capable of giving me these sort of insights. I do read a lot of Wikipedia articles and other sites when looking into math, but just reading words and starting at an equation simply cannot replicate the sort of insights I make during the course of a video such as this.
I'm not an expert in this area, but my understanding is that elliptic curves are used in cryptography because they have a group operation x.y -> z which is easy to compute in one direction, but not in the other. The connection between elliptic curves and modular forms is not relevant to the use of elliptic curves in cryptography. (Wikipedia link: en.wikipedia.org/wiki/One-way_function#:~:text=An%20elliptic%20curve,compute%20k.)
As someone currently studying arithmetic geometry and proof of Fermat's last theorem, this is simply fantastic. You really captured some of the special meaning of the Eichler-Shimura theorem and modularity simply, and in just 10 minutes no less! I'm astonished!
Thank you! Another gorgeous video. Your content is always really well done, and strikes a really, really impressive balance between parsimony and depth of insight. Please keep up the great work! Looking forward to the next one!
Thanks Dan! Hope you enjoyed the video :)
@@Aleph0 hello sir ,i want to approach you please tell email......
I want to become a mathematician ...🙏🙏
Ramanujan is one those few people I just marvel at. He was on a unique plane from all others.
One of the most beautiful and aesthetically pleasing videos in have ever seen. I knew the maths, I knew the journey, but your video, both concisely, logically and especially visually was an absolute delight.
My master's thesis used a lot of modular forms. This was a really nice historical perspective and transitioned so cleanly from the definitions to their impact
ruclips.net/video/nLGYCdEOblI/видео.html
1 year ago yt recommended me the video about derivatives, it's still hard for me but it sparked my interest in maths. Great job on the videos. I hope more people will see them in the future.
I love how you take these very deep technical ideas and show me just enough of them to go wow!
This has hot to be one of the all time best math videos I have ever seen.This is the first time I have been able to understand the connection between modular forms and elliptic curves
Yes, I knew that RSA was connected to modular forms, and that elliptic curves are commonly used in cryptography, but I wasn't aware that they are both connected through modular forms.
I like how you show the genius mathematical intuition of these peoples (even if these ideas took them years to conceptualise).
Thank you for your videos!
That was so freaking beautiful I'm astonished. Thank you so much for making this video. Wow, just wow.
The entirely unintuitive triumph at 6:12 literally made me shout in disbelief. It’s astounding that this isn’t an elaborate gag. Thanks for the great warm-up video.
At this point of time, I think I can safely state that this channel is (at least in my opinion) one of the best mathematics related channels on RUclips.
Btw awesome video!
Wonderful, simple, and intuitive introduction to FLT and modular forms. Your videos never disappoint!
Very nice! Your videos are very engaging. I have decided to support you in Patreon. Best wishes for you and your channel!
Very simplified, but accurate. I hope these videos will stimulate some people to take up the so called hard maths. Very rewarding. Looking forward to more..
Thank you! This gives a nice non-technical intro into a seemingly difficult subject. I'm tempted to redo some of the calculations that you presented.
Hey, just wanted to say that you’re page is freaking awesome. Keep up the great work. That is all.
I've always wanted a video that gave a little insight into modular forms that didn't immediately soar above my head. They always come up in all sorts of places and nobody ever even tries to give the faintest intutition...
Thank you.
I'm actually doing my Ph.D. on a somewhat related topic (transfer operator techniques on hyperbolic surfaces, which can be linked to automorphic forms of which modular forms are a subset. There they appear as building blocks for eigenfunctions of the Laplacian). My impression as to why nobody seems to bothers to give an intuition about modular forms is that nobody really has one. People study them because they are quite peculiar things and seem to hint on profound hidden meaning whenever given the slightest of chances to do so. But apart from that, as far as I know, they remain quite mysterious things. They're like the Dark Matter of math.
@@lonestarr1490 They certainly are mysterious! The closest intuition I've found to modular forms being "natural" objects is that they are sections of line bundles on modular curves, and modular curves are moduli spaces of elliptic curves (which are definitely natural objects!). The way we study spaces in algebraic geometry is by understanding the functions that live on them, i.e. sections of line bundles (or vector bundles, more general sheaves etc.) so in this sense modular forms arise "naturally". Why this in turn has so many applications to complex analysis, hyperbolic geometry, sphere packings, quantum groups, etc. I have no idea, and I have to agree with your perspective that they're just the "Dark Matter of math"
It was a concise explanation that provided me with a high level overview of the subject. Thanks.
P
ruclips.net/video/nLGYCdEOblI/видео.html
You uploaded after a long time!
Your integration(pun intended) of history, small facts and you not fearing of losing viewers because of big maths is really refreshing. We need ofc those dumbed down channels but we also need this, this is like a proper documentary made by BBC that shows the history of a place, showing it's founders and stuff or an Animal planet show etc.
Such great job
One of the most interesting channels.
Your channel won’t stop getting better. No other channel makes videos like yours, and any one that tried would be facing an uphill battle to compete with you.
This is very, very, very, very, very, very, very, very, very, very, very, very good!!!
Very nice overview of a rather complicated and daunting relationship. I look forward to your self-study course on elliptic curves and cryptography.
This is breathtakingly amazing! Thank you!!!
I love your detailed and visually appealing explanations. Great job!
So f(q) is like a very sophisticated generating function. Amazing!
There is actually another connection between a lattice and number theory. If you define a linear, recursive congruence of the form r_{i + 1} = (a*r_i + c) mod M, where a, c and M are constants, then plot successive elements of the sequence as points on a plane (x, y) = (r_i, r_{i +1}) you will see a lattice. If you choose c=1, a=4, M=9, and a starting value of r_0 = 3 you get the sequence 3, 4, 8, 6, 7, 2, 0, 1, 5, 3... and so on repeating. You then will have to plot (3, 4), (4, 8), (8, 6)... and so on. This example is taken from Computational Physics Problem solving with Python by Rubin H. Landau Manuel J. Páez
, and Cristian C. Bordeianu, chapter 4.2.1. It is presented as the linear congruent method for finding pseudo-random numbers. In the book it is given as a bad example of one, considering that the series repeats quite early. I do not know how you would relate this further with elliptic curves, nor if it has been done already.
I love your channel.
I love your channel.
I love your channel.
I don't understand the maths, but I sort of get this from your wonderful video!
Beautiful video, as always!
That was beautiful. Thank you for your videos.
This is a great video. I believe a lot of patters and corollaries we see in mathematical disciplines has to do with the nature of computation. Any multiplication or division operation in a non binary numbering system generates a symmetrical form of operations and ultimately creates prime numbers. It will be cool when people start to crack some of the toughest problems in the fields. It's good to see that curiosity isn't dead!
I'm curious where the 24th power in the first equation comes from. I know that specific equation only serves as background to help introduce the problem, but whenever I see a specific equation with seemingly random constants in them I always wonder where the constant came from.
It's a very important 24 :). The simplest definition of the modular discriminant function (the delta) is delta = g2^3 - 27g3^2, where g2 and g3 are essentially the coefficients of the elliptic curve attached to a modular form (or lattice). Then, g2 and g3 are essentially Eisenstein series; g2 is a sum of -4th powers, and g3 is a sum of -6th powers. So delta ends up being something like a sum of 12th powers. Those are 12th powers of tau, a variable in the upper-half plane; moving tau to q (the "nome") in the unit circle throws in another power of 2. Thus, an overall power of 24. (There's much more, like that delta is the first nontrivial *cusp* modular form, and appears with weight 12, which then becomes 24 after converting tau to the nome. And I think the 24 can also be traced back to the special lattices formed by {1,i}, which gives a factor of 4, and {1, e^{2pi/3}}, which gives a factor of 6. Then 4*6 = 24. I think 4 and 6 are also the nontrivial conductors of imaginary quadratic fields, or something like that.)
@@pinkalgebra Thank you for the easy to understand explanation! I was also wondering why in the world 24 of all numbers was used, but this makes a lot of sense.
@@pinkalgebra well my last comment got deleted for linking to, of all things, a stackexchange post about numerically solving this stuff in Java...
A lot of that initially went over my head, and I had to do a ton of digging through Wikipedia and other sites until I managed to piece together a semblance of understanding, but I'm beginning to understand what you're saying.
I really appreciate the reply, as it has led me to quite a few interesting things to read, and has definitely helped point me towards a better poverty understanding of the topic.
I mentioned in another comment here I'm self taught beyond pre-calc and the power rule for solving derivatives (and I have a tendency to not follow the standard order school trends to teach things, often diving into the deep end and trying to piece together an understanding of things). So while I still don't fully "get" the math, I can see where the numbers come from a bit better.
Very nice historically grounded exposition. RUclips puzzles me given that someone who is conversant with the fine details of many theorems mentioned is likely watching and yet there is in general no middle ground between light popularization and heavy research talks on this platform. Mathematicians are shy creatures, perhaps this furnishes an explanation, not to mention that RUclips is somewhat of an ad riddled commodification machine that seems almost designed to repel serious mathematical discourse. Perhaps we can aim for a sort of digital common room in the distant future. Science shouldn't be confined to the limits of earshot.
!Yes
The legend is back 🔥🙌
beautiful observation by sir srinivasa ramanujan 🙏
Great video indeed. Just watched a Terence Tao video few hours ago. I am lacking complex geometry and various maps knowledge. Elliptic curves occur in other number theory domains, for example congruent numbers, and even theta-congruent numbers, where the elliptic curve has a specific form. I am working on mathematical billiards and they occur there too.
You've done it again man, amazing explanation
well made video man,i really enjoyed it .
This was a great video. On a side note which I think is super cool, I saw the picture of Conrad at the end and could have sworn it was my old Calculus 2 professor (Keith Conrad). Turns out they’re brothers, crazy. They’re nearly identical
Great explanation! I'm a bit confused at 5:43 though. For p = 2 how could there be 5 solutions if there are only 2^2 = 4 possible ordered pairs mod 2? Aren't the only solutions (0,0) and (1,0) since the right-hand side is always even?
I was looking at that too
I'm also confused about this :/
A few seconds before this, he also says there are 4 solutions for p=5. Then here in the chart he says 5 solutions for p=5. I’m lost.
I think the right elliptic curve is y^2+y =x^3-x (you may modified to make it the normal form) and the number of solutions should be modified to be +1(th infinity point is included)
Correction: y^2+y =x^3-x^2 should be the right curve
I was actually thinking about this
i love number thoery, subbed.
I was hoping you’d go into HOW these conjectures were proven, but everything else was really well explained! Good work 👍
5:53 Isn't the next prime 11 and not 13?
It could be that it's a modular form of level divisible by 11. The theorem excludes a finite set of primes, namely those dividing the "level" and the auxiliary prime of your Galois representation. However, the pattern holds for all remaining primes.
Indeed, it's the unique cusp form of weight 2 and level 11. www.lmfdb.org/ModularForm/GL2/Q/holomorphic/11/2/a/
See another similar qn here for a very thorough answer.
@3:45 why x=sin t and y=cos t when usually we say x = cos t and y = sin t (in this case it’s interchangeable)
Cos y not?
@@imnimbusy2885doesn’t the lattice not match?
So, a few questions that immediately spring to mind: Of course for any elliptic curve you can just make a list of all solutions mod all the p's and then write an f(q) with those coefficients, so...
1) what has to be done to prove that this is modular (like, is being modular any more special than looking like what you showed)
2) What has to be done to show that f(q), integrated over all the 'special arcs', gives a lattice which can be fitted periodically by a set of functions (x and y) satisfying the original elliptic curve ( y = f(x)) that it all came from?
In other words, are these things comparably difficult to prove as all the other difficult things that you mentioned people were trying to prove here, or have I missed some 'obvious' logic?
Does multiplicativity only hold for coprime exponents? What's the meaning of the prime power coefficients?
(-2q²)(-2q²) ≠ +2q⁴
(-q³)(-q³) ≠ -2q⁹
EDIT: probably also related to the number of solutions in a ring. That would also explain the multiplicativity, comes from ring factorization. But which ring? Cyclic? Galois field?
Awesome video! Within the next year, I'll be explaining FLT in detail on my channel over a very long series of videos, for those interested.
I may have to check your videos out. I've been wondering what the connection to the torus is, so hopefully your second video will be able to provide that insight.
@@Bobbias there's some wild stuff going on there. For FLT, I'll be doing a series of lectures much like for the Congruent Number Problem, but then I'll also turn everything into a much longer series of refined, shorter videos (10-20 min each) with a little higher production quality correcting any mistakes from the lectures.
My legend is back :)
How can there be more than 2^2=4 solutions mod 2 at 5:45?
I am not sure if this is correct, but note that an elliptic curve includes the point at infinity - it works in projective coordinates.
Thanks!
@@garethma7734 You are correct, but additionally the equation is singular when reduced mod 2, 3, or 11, so there's some special-case handling needed for those primes.
Quick question: At 5:11, Aleph 0 states that a common question, when given an equation, is "how many solutions are there mod p," where p is a prime number. However, at 5:59, when Aleph 0 goes on to list p and the number of solutions to mod p, they do not include 11 as one of the p values. Was this just a simple mistake or does the number 11 not satisfy the condition that 1 + p- (# of solutions mod p) is represented in the coefficient of the modular form?
That's a great observation given how often it's overlooked in various expositions , which by the way makes sense given the technical nature of the question.
If you read the description of this video at some point aleph0 mentions how those lattice points are obtained by paths on that unit disc which are sort of invariant under the action of something called Gamma_0(11). Notice how 11 shows up here. This number 11 associated to a Modular form is called Level of a modular form. It appears in the very definition.
Moreover when we match our modular form with the elliptic curve what we are really looking for is to have an analogous notion of level on the side of elliptic curve as well so that they both the values of levels match. In the world of Elliptic Curves that notion already existed and is called " Conductor ". It measures the reduction type of your elliptic curve what that means is , for all prime numbers p dividing the conductor , the equation you get after reducing your Elliptic Curve mod p will no longer be non-singular and what that means technically is that the tools you have to measure number of primes modulo p on a curve no longer work for that non-singular curve.
In our case since Level of modular form is 11 , the conductor of corresponding elliptic curve would be 11 as well , and the prime dividing this conductor is 11 itself , the so called ' bad prime ' and so reducing your elliptic curve mod 11 would give you a non singular curve and so you would want to avoid it.
@@pursuingstacks Thank you so much for the thorough answer!
P=11 actually is the starting solution node for FLT. The cycle repeats at P=10011, as on a torus . The solution using a secret coefficient node is e^(pi+8), such that the distance between exponents is 2807/1579=1.7777074, starting from p=11
sublime clarity, thank you
This is a good math channel
Ramanujan's q series is one of the Partition functions.
For no particula reason? When harvest starships will come back to Earth and found Ramanujan's work and decide to save our species, you'll get the reason.
Great video, as usual! I have been fascinated with modular forms since I started reading about them in popular summaries of FLT and Wiles' proof. What are the steps towards getting deeply acquainted with them mathematically? I imagine an undergraduate course in Complex Analysis would be the starting place, wouldn't it?
Some complex analysis is required, but not that much. A fair amount of number theory is required. I suggest checking out Tom Apostol's textbook on modular forms. I believe he gives you the pieces of number theory and complex analysis that you need, and you can freely skip over the more technical details of stuff like contour integration which is required for some proofs. The main skill required may be being good at manipulating infinite series and infinite products. Good luck and have fun!
This video just blew my brain
My toxic trait is believing I could’ve solved the Taniyama-Shimura conjecture by myself
@ 5:54 how are there 5 solutions mod(2)?? There are only 4 possibilities?
See another same qn. It involves points at +/- infinity.
To me was very educative. Thanks a lot
Your content is really unique. Thank you so much for sharing it. 🙂
"A reasonable reaction is that this seems very unlikely." Yeah of course. I was just about to say that.
As someone mostly self taught beyond pre-calc, just seeing x = sin theta, y = cos theta, and x^2 + y^2 = 1 helped me see the connection between sin/cos and the unit circle in another way of never noticed.
If x and y are rational, then you multiply though by the denominator and you get a right triangle with integer coefficients
Brilliant video. Make more on the same topic. Thank you
Such a great topic! I've been just thinking about how you can tell the GCD of two numbers by looking at where two sine waves with the period of those two numbers meet at 0!
Wow… studying this closely.
Can you make a video on the symmetric product and algebra?
Beautiful
Thank you so much for your videos
Your videos are really good!
every time i see the abbreviation FLT the first thing that comes to mind is Fermat's little theorem
NT guy
beautiful 💕!
but how can we find the no.of solutions mod p 🤔?
Dear Aleph,
It's my impression or you removed some content of your channel? If yes, could you tell which videos are not available anymore?
How can there be 5 solutions mod 2? There are only 4 possibilities for the pair (x, y), namely (0,0),(1,0),(0,1),(1,1). I'm confused...
Good point :) The solutions are viewed in projective space, so there’s also a “point at infinity”.
I agree @omri, and also, y = 1 does not work mod 2, since the equation reduces to y^2 = 0 mod 2.
Elliptic curves are dual to modular forms.
Thank you, great video! Please make more of them : )
I've seen videos by Richard Borcherds about this stuff. I didn't understand those either.
That has got to be one of the least satisfying ends to an aleph null video ever. Where can I find more?
Also, if I may ask a few questions, what is your specific area of study? What schools did you attend? And where do you find your articles and sources for your videos? Also how did you get so good at explaining these harder or more advanced topics?
There are numerous textbooks on modular forms and elliptic curves. I started with Tom Apostol's number theory books, which has modular forms as a subject, but doesn't really do modular forms. There's a book by Lawrence Washington on elliptic curves that I'm finishing up now, which is probably my favorite intro to elliptic curves, and does only a little bit on modular forms (though a lot on the Weierstrass p-function). So those are some places to find more. You can also just read the Wikipedia and/or Mathworld articles on the things mentioned. Have fun!
@@pinkalgebra thank you very much!
Love from India ❤️😍♥️♥️😍😍❤️
Nice video, given I know nothing about this, I could follow it pretty well. I just need to clarify, what exactly is a modular form? It wasn't explained so clearly.
Clear and concise! I'm a new subscriber :) Just curious, what is the 'so what?' of this finding? I'm sure it's a broad impact connecting modular forms to elliptic curves, and it's connection to Fermats last theorum, but what would be a prominent example of the use of this discovery?
The Weierstrass P function is a thing of beauty. Unfortunately, the Weierstrass pee function didn't operate nearly as efficiently. HEY YOOO...
Thank you for this great video.
Awesome video!
Incredible!
Excellent 👏🏻. And thanks for bringing this up.
y0oooo wwtf this channel is fire
Amazing as always ❤
Are there any relations or info regarding the intermediate coefficients?
For example at 0:43 you say -24*252 = -6048.
Then soon after, *skipping* -1472 you say -24*4830 = -115920.
And that's b/c the place/index of the coefficients, ordered, need to be coprime.
So what about other coefficients?
Yaaay new video!
The paper "On the modularity of elliptic curves over Q: wild 3-adic exercises" is about 100 pages. It will make good sense when it can be written easily over a few pages.
Modular forms and elliptic curves
Infinite phi revolving around infinite parallels
Fractals of infinite realities
each cascading, gliding
In an infinite wheel
Tell me the true nature of my reality!
These are the verses that made me want to become a mathematician.
Great video! But I still don't understand how you connect the delta(q) (e.g. delta(q) = q[(1-q)(1-q^2)(1-q^3)...]^24), the resulting lattice to elliptic curve equation then the function f(q) = q - 2q^2 - q^3 +2q^4+ q^5+2q^6-2q^7-2q^9-2q^10+q^11+4q^13....). Where does the last f(q) polynomial expansion come from relative to the first? I get that the solutions (1+p-) (# of solutions mod p) reveal the same coefficients as this equation, but don't grok where it comes from as one steps through your video. Seems to pop out of nowhere? (e.g. the f(q))