Put . These sigma-fields increase, and their union generates the continuous Brownian path. Let be the polygonal enhancement. Its coordinates are determined by its displacement and antisymmetric area ; the symmetric second level is .
The given conditional-expectation identities imply
For the second assertion it is important to use area coordinates rather than asserting that the whole tensor signature is a conditional expectation. Write . The local off-diagonal second integral is
Conditional on their separate dyadic values, the two independent Brownian components remain independent. Thus the conditional expectation of the product in the last term is . The assumed identity for gives the polygonal entry. The reverse entry follows by integration by parts, since the product of the two component increments also has the product of their conditional means. Taking the antisymmetric part proves the area identity. Diagonal second levels are recovered from the geometric identity, rather than from a conditional-expectation claim about squares.
For each fixed dyadic time , the first-level and area conditional expectations are uniformly integrable martingales in . Their targets are measurable with respect to : Brownian area is a limit of stochastic integral sums determined by the path. The martingale convergence theorem gives convergence to and , respectively. The continuous group coordinate formulas then give
at all dyadic times on a common event of probability one.
For the uniform Hölder estimate, fix and put . The given all-order integrability and the group norm comparison imply
Use the conditional-expectation formulas, the conditional Jensen inequality, and . The group norm comparison yields
First take all dyadic pairs, a countable collection, and then extend the inequality to all by continuity. Hence the Brownian rough path martingale Hölder bound is
For any , the Doob Lp maximal inequality gives
Therefore
To upgrade convergence from dyadic times to every time simultaneously, apply this bound at one positive exponent . The polygonal enhancements are then equicontinuous in the group metric, with a common finite random constant. For a finite dyadic time net, convergence holds at every net point. The triangle inequality bounds the errors at the remaining times by the two Hölder errors to the nearest net point and the error at that point. First let , and then let the net mesh tend to zero. This proves uniform convergence, and in particular convergence for every on one event of probability one.
There is also convergence in each weaker Hölder rough path metric. Uniform convergence of the group coordinates gives uniform convergence of their levelwise increments through the Chen formulas. If , the uniform bounds bound the level- increment differences by a random constant times , while their unweighted suprema tend to zero. Interpolating these two bounds proves convergence with denominators , for . This establishes the nested polygonal approximation of Brownian rough paths in the topology needed for the support argument.
For Brownian motion starting at zero, define its enhanced Brownian motion by the Stratonovich second level:
The associated group-valued path is . The Stratonovich chain rule gives
Consequently its symmetric second level is and the remaining antisymmetric part is Lévy area. Splitting the integral at proves the Chen identity, so these increments lie in and are the relative increments of .
The group norm for the Carnot-Carathéodory distance is comparable to . The planar construction in Question 2 extends to dimension : prescribe every coordinate-plane area with a closed rectangle, concatenate these central-area loops, and finish with the displacement segment. The converse bound follows from length bounds on displacement and on each area integral. Thus the comparison constants depend only on .
Brownian scaling and stationarity of increments imply that has the distribution of . The permitted moment estimates for Lévy area, together with Gaussian moments of Brownian displacement, therefore give, for every finite ,
This also verifies the hypothesis of the metric-valued Kolmogorov continuity theorem. We give the dyadic argument to obtain a simultaneous version explicitly.
Fix and choose so large that . The Markov inequality and a union bound give
The right-hand side is summable. By the Borel-Cantelli first lemma, almost surely there is a finite random constant bounding every adjacent dyadic increment by , including the finitely many exceptional levels.
Decompose any interval with dyadic endpoints into successive dyadic intervals, with at most two intervals of each size and no size exceeding its length. The metric triangle inequality and the sum of the geometric series give
for all dyadic . The finite-dimensional Carnot group is complete, so the path extends uniquely from dyadic times to a continuous -Hölder path. At each fixed time the original stochastic enhancement and its values at approaching dyadic times converge in probability by the moment estimate. The extension has the same limit almost surely, hence is a modification of the original enhancement.
Perform this construction on one countable sequence of exponents increasing to . The extensions agree on the dense set of dyadic times, so they give one modification that is Hölder at every exponent below . On , a -Hölder estimate implies the same estimate at every . Therefore
Let be a weak geometric p-rough path, , with increments indexed by words of length at most . Choose a control function for rough paths with . We use the standard sufficient vector-field hypothesis , , with globally bounded regularity norms; smooth bounded fields are a special case. Regard as the differential operator , and put .
A pathlevel solution of a rough differential equation is a continuous state path , with prescribed , satisfying a local expansion
for some on sufficiently small control intervals. One can take under the stated regularity. This definition specifies the state path, without requiring its full rough-path enhancement as additional data. For , the displayed expansion reads
Repeated indices are summed.
We use the following discrete form of the Davie lemma. For a finite ordered grid and a continuous superadditive control, suppose nonnegative errors vanish on adjacent grid points and, for every grid triple , satisfy
There are and , depending on but not on the grid, such that whenever . This is the discrete higher-order error estimate: splitting close to half the control size makes the two -power errors contract; a crossing single grid interval contributes zero adjacent error.
Define the local Euler map
For small it has Lipschitz constant at most . Taylor expansion, the Chen identity and the shuffle identity for iterated integrals give the composition estimate
Here is the cancellation underlying the estimate. Assign degree to a level- increment. Expanding each and grouping terms through degree gives the differential-operator coefficients multiplying . Products of coefficients from the first increment are combined by the shuffle identity, which is the iterated Leibniz rule. Chen's identity says that the sum of these products is . Thus every term of degree at most agrees with the corresponding term in . The remaining Taylor terms have control order at least ; bounded norms make the estimate uniform in . For smooth fields the remainder has order .
On a partition , construct the Euler sequence by . For grid endpoints let
It vanishes on adjacent points. Insert into the next local map. The Lipschitz and composition estimates give
The Davie lemma therefore yields a bound independent of on small control intervals. Together with the Euler expansion this gives . A finite subdivision into small-control intervals gives a uniform global bound as well.
Interpolate each Euler sequence linearly in time, and take partitions with mesh tending to zero. Continuity of the control makes these paths uniformly equicontinuous: the grid estimates apply to the interval enlarged to its neighboring grid endpoints, and the enlargement tends to zero uniformly with the mesh. The paths are uniformly bounded. The Arzelà-Ascoli theorem supplies a uniformly convergent subsequence with continuous limit .
For arbitrary in a small control interval, choose neighboring grid endpoints converging to . Continuity of , the coefficient functions and the control lets us pass to the limit in the grid error estimate. We obtain the defining expansion with , and . This constructs a pathlevel RDE solution. No continuity theorem for the RDE solution map was used to obtain existence.
Use the truncated tensor algebra , with truncated concatenation product. The step- path signature is
These iterated integrals of a path exist as Riemann-Stieltjes integrals, since a Lipschitz continuous path on a compact interval has bounded variation. The degree-zero coordinate is , and the degree-one coordinate is .
Let and write for its degree- coordinate. Integrating first over all variables except the last gives
Thus for . If is a standard basis vector, define the linear vector field , with degrees above discarded. The component identities are precisely the controlled ordinary differential equation
For a Lipschitz driver this may also be written almost everywhere. Recursive integration of its levels proves uniqueness and recovers the signature formula. The Chen identity follows by splitting each integration simplex at an intermediate time.
To correct every pattern of errors by nearest-codeword decoding, it suffices to have minimum Hamming distance : the corresponding radius- Hamming balls are disjoint. For , the relative distance tends to , so the asymptotic Gilbert–Varshamov bound guarantees code rate at least in the limit. Therefore its immediate strict-rate guarantee is
The binary entropy is strictly increasing from zero to one on , since on . Let denote the inverse on this interval. The asymptotic condition becomes
There is also a finite-length boundary refinement for the precise requirement of correcting errors. If , then each weight for is at least . Expanding gives the entropy bound for a Hamming ball
The case follows directly from . Taking and using monotonicity of the binary entropy, the Gilbert–Varshamov bound yields
Thus the worst-case correction guarantee from the Gilbert–Varshamov bound even includes for the exact radius , with the number of messages rounded to an integer when necessary. No positive-rate guarantee of this form is obtained for ; one must not extend the increasing-entropy calculation past . Failure of this sufficient bound does not rule out better codes.
By contrast, the Shannon second coding theorem gives vanishing error probability on the binary symmetric channel whenever
The factor of two comes from demanding disjoint correction balls for every error pattern of a prescribed weight. This is a worst-case distance requirement, and the Gilbert–Varshamov bound is itself only a sufficient existence bound. The Shannon second coding theorem controls the probability of decoding failure for the random channel; rare bad patterns may remain.
Finally, the number of actual errors has binomial distribution with mean . Correcting exactly its mean does not itself imply error probability tending to zero. To get that conclusion from the distance construction when , choose with and , and correct errors. The weak law of large numbers then makes the probability of exceeding the correction radius tend to zero.
Figure 1.
Binary symmetric channel capacity and Gilbert–Varshamov worst-case correction rate
. Details.
The Shannon second coding theorem says that a memoryless channel permits transmission at any fixed code rate with error probability tending to zero as block length tends to infinity. Conversely, a sequence of codes with error probability tending to zero cannot have a limiting rate above . The channel capacity, in bits per use, is
For a finite discrete memoryless channel with transition probabilities , the supremum is a maximum and its mutual information is
For a general memoryless channel, mutual information is the relative entropy of the joint input-output law with respect to the product of its marginal laws; the same supremum is taken over the admissible input laws.
For the binary symmetric channel, write with independent noise having Bernoulli distribution of parameter . Given either input bit, the output conditional entropy is . Thus
An input with uniform distribution makes the output have uniform distribution, attaining . The binary symmetric channel capacity is therefore
Write for the largest number of codewords in a binary length- code with minimum Hamming distance at least . The volume of a Hamming ball does not depend on its centre and is
For the Hamming bound, put . The radius- Hamming balls around distinct codewords are disjoint: a point belonging to two such balls would, by the triangle inequality for Hamming distance, put their centres at distance at most . Counting points in the binary cube therefore gives
For the Gilbert–Varshamov bound, choose a code maximal under inclusion among those with minimum Hamming distance at least , for example by successively adding any allowable word. Its radius- Hamming balls cover the cube. Otherwise an uncovered word would be at distance at least from every codeword and could be added, contradicting maximality. Thus
The lower bound is an existence assertion about a suitable code, while the upper bound holds for every such code. Together they give
Let be the binary entropy. For ,
The assumed Hamming ball volume exponent applies also to any integer radii with . Indeed, for each small these radii eventually lie between and . monotonicity of sandwiches its normalized logarithm between limits and ; continuity then lets tend to zero. Consequently the asymptotic Gilbert–Varshamov bound and asymptotic Hamming bound have respective exponential scales
Equivalently, the optimal code rate satisfies
These are bounds on the rate, not a claim that the optimal rate has a known limit or equals either endpoint.
For , the Poisson distribution gives
This information entropy is finite: for integer , while .
Fix . Take independent random variables and . Their convolution is
by the binomial theorem. Thus has Poisson distribution with parameter . Conditioning on only translates , so . The entropy monotonicity under independent addition yields .
In fact the inequality is strict. The covariance is positive, so and are not independent random variables. Their information entropies are finite, and the equality criterion from part (a) excludes equality. Therefore the Poisson entropy monotonicity gives the stronger conclusion
Introduce a selector with Bernoulli distribution of parameter , and construct by taking its conditional law given to be , and its conditional law given to be . No prescribed joint law of is needed. Summing over the selector gives the desired mixture probability mass function. The conditional entropy is
Since conditioning cannot increase information entropy,
This proves the concavity of information entropy. The argument also covers infinite information entropies with the convention that a zero mixing weight contributes zero. At or , the mixture is the corresponding original law and equality is immediate.
Use base-two logarithms and the convention . Write and for . The conditional entropy is the average of the information entropies of these conditional laws:
The careful form of Gibbs inequality is that, for two probability mass functions on a common countable set,
where a term with gives . Equality holds exactly when . This is the nonnegativity of relative entropy.
First suppose . For every of positive probability, implies . Averaging the cross-entropies gives
All the summands here are nonnegative, so interchanging the sums is valid, and each conditional cross-entropy is finite. Applying Gibbs inequality to each shows that is at most its cross-entropy. Subtracting and averaging therefore gives
Every term on the right is nonnegative. The equality in the entropy conditioning inequality is thus equivalent to for all , which is exactly factorization of the joint probability mass function. Consequently
For arbitrary countable discrete random variables, the inequality remains valid in : the case is immediate. The finite-entropy condition is necessary for the usual equality characterization. For example, let for , with the normalizing constant, and let be an independent fair bit. The series defining the probability mass function converges, whereas its information entropy diverges like . Then and are dependent, but . In the infinite case, equality means precisely that the conditional entropy is also infinite; it does not imply independent random variables.
Conjugation is an automorphism of sending to . It induces the bijection
between their left coset sets; conjugation by gives the inverse. Therefore
This holds whether or not is normal, and whether the index is finite or infinite.
For an arbitrary subgroup , the full inverse image of its image is , not generally . In fact, exactly when for some . Since is normal, is a subgroup. By the subgroup correspondence for a surjective group homomorphism,
Multiplicativity of finite index gives
The map bijects the left cosets of in with those of in , since . Consequently
This need not equal . For example, take to be reduction modulo and . Then , but has index one. If contains , however, the denominator is one and
For finite , equality holds exactly when .
For a finitely generated group , every quotient group is generated by the images of the . Thus finite generation passes to quotients.
For a group extension , suppose and are finitely generated. Choose finite group generators of and lifts to of finite group generators of . For any , a word in the lifts has the same image as in , so their difference belongs to . Thus the two finite sets together generate , proving preservation under extensions.
For a subgroup of finite index, choose representatives of the right cosets , with representing , and a finite symmetric generating set for . Write for the chosen representative of . The finitely many Schreier generators
belong to . If , set , so . Then
Each factor is a listed generator because . Therefore is finitely generated. This proves Schreier's lemma directly.
An infinite locally finite group is
Its elements are binary sequences of finite support, with coordinatewise addition. It is infinite because the unit vectors are distinct. Any finite collection of elements has its supports contained in one finite set of coordinates, and the subgroup it generates lies in the corresponding finite product of copies of . Hence every finitely generated subgroup is finite.
Local finiteness passes to subgroups because a finitely generated subgroup of a subgroup is also one of the ambient group. It passes to quotients as well: lift finitely many group generators of a subgroup in the quotient; the subgroup generated by their lifts is finite, and its image is exactly the desired subgroup.
For local finiteness is closed under extensions, suppose and are locally finite, and take a finitely generated subgroup . Its image is finitely generated, hence finite. The map has kernel , so
The finite-index result already proved makes finitely generated. Local finiteness of makes it finite, and then
Since was arbitrary, is locally finite. Thus finite generation is preserved by quotients, extensions and finite-index subgroups; local finiteness is preserved by subgroups, quotients and extensions.
Arithmetic heights measure arithmetic complexity. For a reduced rational number , its multiplicative height is and its logarithmic height is . Over a number field , use the absolute projective height
Normalize the absolute values on a field by the product formula. That formula makes arithmetic height independent of the chosen homogeneous representative, and the normalization makes it independent of enlarging . The Northcott theorem says that algebraic points of bounded degree and bounded arithmetic height form a finite set. For rational numbers this follows immediately by bounding the numerator and denominator. In a fixed number field, the same finiteness follows by bounding the coefficients of the minimal polynomial using its conjugates and the arithmetic height.
For an elliptic curve put and . The two points above a given show that bounded gives finitely many -rational points. More importantly, multiplication expands this arithmetic height. To see the mechanism, choose a short Weierstrass equation of an elliptic curve over . Doubling gives
After homogenization the numerator and denominator are quartics with no common zero. At infinity the numerator is nonzero; at a root of the cubic the numerator becomes , nonzero by nonsingularity. For a degree- morphism of the projective line, fixed homogeneous coefficients give the upper arithmetic height bound . A nonzero resultant supplies the reverse bound: at each place it bounds the common cancellation of the two homogeneous values, with a nontrivial constant needed at only finitely many places. Applying these bounds to duplication gives a constant independent of with
This includes points mapping to , by the projective formulation.
The canonical height removes the bounded errors. With the conventional factor one half, define
The successive terms before multiplying by one half differ by at most . The limit exists, and summing this geometric bound gives
Hence bounded canonical height still gives finitely many points over the fixed number field.
To obtain the pairing, the quadratic nature comes from a uniform arithmetic height parallelogram estimate, not just from doubling. If and , the unordered coordinates are the roots of the binary quadratic with coefficients
These formulas follow by adding and multiplying the two chord expressions with slopes and . Their homogenizations have bidegree and no simultaneous zero. Off the diagonal ; on the finite diagonal , and for , so simultaneous vanishing would contradict nonsingularity. At the double point at infinity the homogeneous is nonzero. Local coefficient norms, or the corresponding resultant bounds, therefore give
The arithmetic height of a binary quadratic's coefficient vector differs by from the sum of the heights of its two roots. At finite places this follows from multiplicativity of the maximum coefficient norm of products; at infinite places the fixed-degree norms are comparable. Thus
Apply this to , divide by and take limits. The errors vanish and give the exact height parallelogram identity. Polarization consequently makes
a symmetric bilinear pairing. Nonnegativity of for integers , and density of rational ratios, give the Cauchy-Schwarz inequality. In particular . Also for integers . Torsion points of an elliptic curve have canonical height zero because their doubling orbit is finite; conversely, if a point has canonical height zero, all its integer multiples have bounded naive height, and finiteness forces two multiples to coincide. Thus canonical height vanishes exactly on torsion.
There is an arithmetic finiteness input before height descent: is finite, the Weak Mordell-Weil theorem. Here is its mechanism. The multiplication exact sequence gives an injective Kummer map : choose a half-point and use the cocycle ; changing changes a coboundary, and a trivial class means a half-point can be chosen over . Let , a Finite Galois extension. Over the torsion module is , so Kummer theory identifies its first Galois cohomology with two copies of .
Only classes unramified outside a fixed finite set occur. Include the primes over , bad-reduction primes and the ramified primes of . At a good prime away from , multiplication by two on the proper smooth elliptic model is finite étale morphism, so the torsor of halves of a point is unramified. In square-class language, valuations outside are even. Such classes form a finite group: write the principal ideal of a representative as a square times an ideal supported on . The ideal class gives an element of the finite two-torsion of the localized ideal class group; after fixing that class, ambiguity is an -unit modulo squares. The -unit theorem makes the latter finite. The group kernel of restriction from to is contained in , which is finite because both the group and module are finite. This proves the weak theorem, without assuming finite generation of .
Finally choose representatives of and put . Every has the form , and
If , this reduces canonical height by more than one half. Iterate until a point of canonical height at most is reached. There are only finitely many such points, by Northcott and the bounded difference of heights. Reconstructing at every stage proves that these finitely many small points and the finitely many representatives generate . Therefore
The canonical pairing is positive definite on the free part. It turns rank computations into an arithmetic lattice problem, supplies independence tests for proposed generators, and bounds searches once a arithmetic height bound is available. Weak finite-quotient finiteness alone would not prove this result: the arithmetic height contraction is the step that converts it into finite generation.
Here is the usual one-dimensional commutative formal group law. Normalize the valuation by , and put . We show explicitly that for any integer ,
This subgroup will have finite index in .
Set . Since terms of total degree at least two, . Its inverse has integral coefficients, say with . Define the formal logarithm over by
Differentiate associativity with respect to at zero. This gives , so
Evaluating at proves the formal identity .
The coefficients satisfy . Thus the series converges on , since for the valuations of its terms are at least . More quantitatively,
Indeed, if then . If , every nonlinear term has valuation strictly greater than . Write . For , use to obtain
The extra valuation is integral and positive because for all .
For any , solve by iterating . This map preserves the complete ball and improves differences by at least one valuation step. The contraction mapping theorem gives a unique solution. Therefore is a bijection of this ball onto itself, and the formal identity, now evaluated in convergent series, makes it a group isomorphism. Multiplication by then identifies the additive target with .
Integral coefficients in and its formal inverse ensure that every is a subgroup. Reduction modulo respects the group law; two parameters represent the same coset of exactly when they are congruent modulo . If the residue field has elements, the quotient consequently has elements. This proves both the finite-index assertion and the claimed additive structure, with an explicit depth that also works in ramified extensions and when .
We can use reduction instead of expanding a very large rational addition formula. Modulo , both and reduce to , which generates the three-element reduced group. The reduction homomorphism therefore gives
The rational point is not any of the three torsion points just found, so it has infinite order. In particular : otherwise multiplication by three would give .
If had integral coordinates, its affine integral representative would reduce to an affine point modulo , not to the point at infinity. This contradicts its reduction to . Thus
Indeed, it is not even integral at . This is a reduction certificate for a nonintegral elliptic point, with the crucial nonidentity condition verified by the torsion calculation.
Use the standard theorem on reduction of torsion points on an elliptic curve: at a prime of good reduction, reduction is injective on torsion whose order is prime to that prime. Reduction at implies that every odd-primary part of the rational torsion subgroup has order dividing . Thus its only possible odd prime is , and its -primary part has order at most . Reduction at injects the -primary part into a group of order , so that part is trivial. No other primary part can occur.
Consequently the torsion subgroup has order at most three. Since has exact order three,
This argument uses only prime-to-residue-characteristic injectivity, and therefore avoids an unjustified claim of full torsion injectivity at the prime .
The primes do not divide the elliptic-curve discriminant , so all three reductions are nonsingular. Count affine solutions and add the point at infinity. At the right side takes values at , while for both values of . This gives two affine points, namely .
At , the right-side values at are . The corresponding numbers of solutions to are , giving five affine points. At , the right-side values at are , with respective solution counts , giving eight affine points. Therefore

Pinned article: Introduction to the OurBigBook Project

Welcome to the OurBigBook Project! Our goal is to create the perfect publishing platform for STEM subjects, and get university-level students to write the best free STEM tutorials ever.
Everyone is welcome to create an account and play with the site: ourbigbook.com/go/register. We belive that students themselves can write amazing tutorials, but teachers are welcome too. You can write about anything you want, it doesn't have to be STEM or even educational. Silly test content is very welcome and you won't be penalized in any way. Just keep it legal!
We have two killer features:
  1. topics: topics group articles by different users with the same title, e.g. here is the topic for the "Fundamental Theorem of Calculus" ourbigbook.com/go/topic/fundamental-theorem-of-calculus
    Articles of different users are sorted by upvote within each article page. This feature is a bit like:
    • a Wikipedia where each user can have their own version of each article
    • a Q&A website like Stack Overflow, where multiple people can give their views on a given topic, and the best ones are sorted by upvote. Except you don't need to wait for someone to ask first, and any topic goes, no matter how narrow or broad
    This feature makes it possible for readers to find better explanations of any topic created by other writers. And it allows writers to create an explanation in a place that readers might actually find it.
    Figure 1.
    Screenshot of the "Derivative" topic page
    . View it live at: ourbigbook.com/go/topic/derivative
  2. local editing: you can store all your personal knowledge base content locally in a plaintext markup format that can be edited locally and published either:
    This way you can be sure that even if OurBigBook.com were to go down one day (which we have no plans to do as it is quite cheap to host!), your content will still be perfectly readable as a static site.
    Figure 5. . You can also edit articles on the Web editor without installing anything locally.
    Video 3.
    Edit locally and publish demo
    . Source. This shows editing OurBigBook Markup and publishing it using the Visual Studio Code extension.
  3. https://raw.githubusercontent.com/ourbigbook/ourbigbook-media/master/feature/x/hilbert-space-arrow.png
  4. Infinitely deep tables of contents:
    Figure 6.
    Dynamic article tree with infinitely deep table of contents
    .
    Descendant pages can also show up as toplevel e.g.: ourbigbook.com/cirosantilli/chordate-subclade
All our software is open source and hosted at: github.com/ourbigbook/ourbigbook
Further documentation can be found at: docs.ourbigbook.com
Feel free to reach our to us for any help or suggestions: docs.ourbigbook.com/#contact