Affine combination Created 2026-10-14 Updated 2026-10-14
An affine combination is a weighted point with . The coefficients can be negative. The sum is meaningful in an affine space without selecting a preferred origin: relative to any point , it means . Affine maps preserve these weighted points. [ref]
For an original example in the Euclidean plane, take , and . Then lies inside their triangle. Its coefficients are nonnegative, making it a convex combination. In contrast, is beyond : it is an affine combination but not a convex combination of these three points. Negative weights allow extrapolation, rather than only interpolation inside the convex hull.
For the affine map , the same weights produce from , and produce from . Thus one can compute a weighted point before transforming the vertices or afterwards, obtaining exactly the same point.
This is useful whenever an object is built from weighted control points: its construction commutes with an affine map, so transforming the control points transforms the constructed object as well. The sum of the weights is the essential condition; no assumption about lengths or angles is needed.
Affine equivariance of a geometric basis Created 2026-10-14 Updated 2026-10-31
A geometry representation with scalar partition of unity basis functions commutes with every affine map of its control points. For , linearity moves through the sum and the sum-to-one identity supplies the translation . Positivity is needed for a convex hull property, but not for this equivariance. The same argument applies to linear subdivision matrices whose rows sum to one. General projective transformations require homogeneous control points and rational dehomogenization rather than transforming Euclidean controls with unchanged weights.
Affine group Created 2026-10-07 Updated 2026-11-05
For a finite-dimensional vector space , the affine group consists of invertible affine maps , where belongs to the general linear group and . Its product is , making it the semidirect product of the translation vector space and its general linear group. On the real line the positive-dilation subgroup is the real affine group used in the adjacent article.
Affine image of a circle 2026-10-20
Let a radius- circle have center , and apply the real affine map . If is invertible, substitute into to obtain the displayed ellipse. Its center is and its semiaxes are times the two singular values of , as follows from the singular value decomposition. A rotation or reflection preserves both semiaxes and hence gives another circle. If has rank one, the image is a segment; rank zero gives one point. This real-plane description also handles maps that stretch real and imaginary coordinates differently and are not holomorphic complex maps.
Affine iteration 2026-10-25
Iterating an affine map gives . A fixed point exists exactly when is in the image of a linear map . If the spectral radius of is below one, its matrix powers tend to zero and every orbit converges to the unique fixed point .
Affine map Created 2026-10-07 Updated 2026-10-14
An affine map combines a linear map with a translation. After choosing coordinate origins, it has the form , , where the matrix is its linear part of an affine map and is a fixed vector. Equivalently, it preserves every affine combination. [ref]
Curves and Surfaces in Geometric Modeling: Theory and Algorithms by Jean Gallier. §2.7, Definition 2.7.1 and linear-map construction, pp. 37–38
An original planar example isIt stretches horizontally, applies a shear mapping, and then translates. The vertices , and become , and . A square grid becomes a grid of parallelograms; straight edges and their weighted points still fit together.
This map is not a linear map with respect to the displayed origins, because rather than zero. It also is not a Euclidean similarity: has length and has length , while the right angle at becomes . Its determinant is , so it doubles planar area. These metric changes explain why affine geometry studies straightness, incidence and weighted points rather than treating every transformation as an isometry.
The linear part is an invertible matrix, giving the concise inverse . More generally an affine map is bijective exactly when its linear part of an affine map is bijective. Between equal finite-dimensional spaces, this means an invertible matrix; the translation cannot repair a loss of rank of a linear map.
Affine maps can collapse lines 2026-10-14
An affine map need not be injective. For , the direction belongs to the kernel of its linear part of an affine map. Consequently the whole line maps to , and every parallel line maps to . These are collapsed lines, not images with a surviving direction. Assertions that an affine map preserves distinct parallel lines require the appropriate noncollapse or injectivity hypothesis.
Affine pressure-tap coordinate calibration 2026-11-05
An affine pressure-tap coordinate calibration converts an instrument coordinate to a physical coordinate through an affine map, . The offset fixes the datum and the nonzero scale fixes the length conversion. Its adequacy is a question about the actual viewing or measuring geometry, rather than a consequence of having fitted two points.
For two distinct instrument readings at known physical references ,An original imaging example takes the leading edge at pixels and , and the trailing edge at pixels and m. Hence m/pixel and m. A tap at pixels is at m, or dimensionless chord position .
If all three features share the same affine map, the dimensionless coordinate is simplyChanging the displayed coordinate to gives readings 230, 1030 and 470 pixels and leaves . This is the exact pressure-tap datum cancellation; it does not assert that the original feature readings are known without uncertainty.
For example, with , the sensitivity coefficients of with respect to the tap, leading-edge and trailing-edge readings areTheir sum is zero, as required for invariance to a common offset. At the example readings they are per pixel. Use their full covariance matrix when propagating uncertainty, because independently attaching three origin errors would destroy the cancellation.[ref]
Guidelines for Evaluating and Expressing the Uncertainty of NIST Measurement Results by Barry N. Taylor and Chris E. Kuyatt. Appendix A, covariance terms in the measurement-model sensitivity law
An affine pressure-tap fiducial check uses an additional independently located mark. At pixels the fitted map predicts m. A measured location m leaves a 2 mm discrepancy to investigate. Two fitted reference points necessarily have zero regression residual; that fact cannot validate the map between them. Perspective, surface curvature, optical distortion and reference-location uncertainty can all matter. A proposed calibration records extra marks across the image and compares residual patterns before using the map for surface pressure integration for airfoil lift.
Bézier curve 2026-10-14
A Bézier curve is a parametric curve specified by control points in a Euclidean space and the Bernstein basis:Here is a binomial coefficient. Each coordinate is a polynomial of polynomial degree at most ; special choices of control points can lower the polynomial degree. The ordered control points form the control polygon. A cubic Bézier curve uses four control points and has the explicit formThe curve passes through and ; the interior control points steer its shape and need not lie on it.[ref]
https://ocw.mit.edu/courses/6-837-computer-graphics-fall-2012/aba8735cf2cf7c71ca5d29b46873309a_MIT6_837F12_Lec01.pdf. MIT 6.837, Bézier Curves and Splines, slides 23–31 and 45–47, 2012
For a worked example, chooseSubstituting these control points givesHere , so neither interior control point, both at height , lies on the curve. De Casteljau's algorithm computes the same point using only successive linear interpolations.
The endpoint tangent vectors follow from the Bézier derivative control polygon:Thus the curve leaves toward and arrives at from . These nonzero derivatives determine the endpoint tangent lines; if an endpoint derivative vanishes, that formula alone does not determine its tangent direction. Endpoint handles are useful when arranging a parametric first-derivative join of Bézier curves.
The parameter is usually not arc length. For this example,Equal steps in therefore do not generally travel equal distances. Distance along the curve is instead the arc length, given by the integral ; its derivative is the norm of the tangent vector and measures the rate of travel.
Two geometric facts make Bézier curves especially convenient for design. First, the Bernstein basis weights are nonnegative and form a partition of unity, since the binomial theorem gives . Hence is a convex combination of the control points and lies in their convex hull.[ref]
https://groups.csail.mit.edu/graphics/classes/6.838/S98/meetings/m15/. MIT 6.838, Bézier Curve and B-Spline Curve, §1.4.1, 1998
The convex hull is the smallest convex set containing the control points, not their axis-aligned bounding box. Here it is a trapezoid, strictly smaller than : the box corner is outside the convex hull. The curve need not fill or touch every edge of either enclosure. The convex hull guarantee concerns ; outside that interval weights can be negative.Second, applying an affine map to every control point gives exactly the same result as applying it to the entire curve. For ,This is affine equivariance of a geometric basis: a translation, rotation or shear mapping can be performed on a handful of control points. For example, transforms the four control points to and transforms the midpoint to , also obtained from the same weights .
These editable handles explain the usefulness of Bézier curves in computer aided geometric design, vector graphics paths and outline fonts: a small collection of points controls an entire smooth stroke, and resizing the design preserves its mathematical shape. Cambridge's computer-graphics notes describe their use in graphic design, typography and PostScript outlines. The industrial context includes Pierre Bézier's work at Renault and Paul de Casteljau's work at Citroën.[ref]
https://www.cl.cam.ac.uk/teaching/1415/CompGraph/cgip.pdf. Neil A. Dodgson and Peter Robinson, Cambridge Computer Graphics and Image Processing, slides 183 and 226–228, 2014
// Target: analysis.bigb
Brouwer fixed-point theorem Created 2026-09-24 Updated 2026-10-14
The Brouwer fixed-point theorem says that every continuous map has a fixed point when is a nonempty compact set and a convex set in a finite-dimensional vector space over the real numbers. Equivalently, take for a finite , with the usual Euclidean space structure: there is some such that . The conclusion guarantees at least one fixed point, without asserting uniqueness or telling us how quickly to find one. [ref]
https://courses.csail.mit.edu/6.5440/fall23/lectures/L23_images.pdf. MIT lecture slides on Brouwer's theorem
The standard pictures are a closed ball, including a closed unit ball, and a simplex. A filled disk qualifies; its bounding circle alone does not. A probability simplex also qualifies even though its dimension is smaller than that of its ambient Euclidean space. The hypotheses have concrete meanings: being a convex set keeps every line segment between allowed points inside , and compactness prevents escape or loss of a limiting point. In Euclidean space, the Heine-Borel theorem identifies compact sets with closed sets that are bounded sets. Crucially, every output must belong to the same as the input.
For a fully computable example, let be the filled unit disk. Apply a counterclockwise quarter-turn rotation, shrink by , and add a translation:This affine map is a continuous map. Because the rotation preserves the Euclidean norm, the triangle inequality gives, for every ,Thus the whole image lies strictly inside . More precisely, it is a disk of radius centred at . This containment check is what makes the Brouwer fixed-point theorem applicable; a formula being a continuous map by itself would not suffice.
The fixed point can now be found by a system of linear equations:Indeed, substitution gives . Its Euclidean norm is , safely inside . The figure shows arrows from sample inputs to their outputs; an arrow disappears at the fixed point because the displacement vanishes there.
Repeated iterations of a map happen to succeed for this example. Starting at and setting gives , , and . Subtracting the fixed point equation yields , so the Euclidean distance to shrinks by exactly each step. This extra fact follows from the contraction mapping property and the Banach fixed-point theorem. It is additional information beyond the Brouwer fixed-point theorem.
The distinction becomes visible with two simpler continuous maps of the same disk. The pure rotation has exactly one fixed point, , since and force both to vanish. Yet iterations of a map starting at here cycle through , , and forever. Their Euclidean distance from stays one, so there is no convergence in a metric space. The reflection has an entire line segment of fixed points, namely . Starting off that line segment produces an alternating pair. Existence, uniqueness, and convergence of iterations of a map are separate questions.
Removing hypotheses changes the answer. Restrict the same quarter-turn rotation to the circle . It remains a continuous map into that compact set, but its only possible fixed point, , has been removed. The circle is not a convex set. Conversely, on the unbounded convex set , the translation has no fixed point; on the open interval , also has none. Neither domain is a compact set. The finite-dimensional closed ball statement should likewise not be extended to arbitrary infinite-dimensional vector spaces, where a closed ball need not be a compact set.
The one-dimensional version explains the connection to the intermediate value theorem. For a continuous map , put . Then and , so the intermediate value theorem gives for some . In higher dimensions, the Brouwer fixed-point theorem supplies a corresponding existence principle when this scalar sign argument is unavailable. Its practical value is ensuring that a proposed equation has a solution once a suitable domain is found; it supplies no general promise of a fast algorithm.
A useful application is positive matrix eigenvector from simplex normalization: rescale the output of a strictly positive matrix to make its entries sum to one. For example, takeOn this probability simplex, has entry sum , so the normalized continuous map is . Both entries are positive and sum to one, keeping the image inside the probability simplex. Its fixed point equation is , hence . Consequently satisfies : a strictly positive eigenvector with eigenvalue . In larger dimensions, the Brouwer fixed-point theorem guarantees the corresponding positive eigenvector without requiring us to solve a system of linear equations in advance. The fixed point expresses preservation of relative weights, rather than preservation of their unnormalized size.
Proofs can be approached through the no-retraction theorem or Sperner's lemma. In the latter approach, increasingly fine triangulations locate approximate fixed points; compactness supplies a convergent subsequence, and continuous maps turn its limit into an exact fixed point. [ref]
https://www.mit.edu/~6.7980/brouwer.html. Daskalakis and Farina, Brouwer and Sperner
A primary historical reference is L. E. J. Brouwer's 1911 paper Über Abbildung von Mannigfaltigkeiten, whose digitized original is hosted by Zenodo under DOI 10.1007/BF01456812. This gives a direct route to the early literature behind the Brouwer fixed-point theorem. [ref]
https://zenodo.org/records/1428286. Brouwer, Über Abbildung von Mannigfaltigkeiten, 1911
Bühlmann model Created 2026-10-06 Updated 2026-11-05
Given a latent risk parameter , observations are independent and identically distributed random variables with conditional expectation and conditional variance . The Bühlmann credibility premium estimates by an affine map of past observations under mean squared error. Its population parameters are , expected process variance , and variance of hypothetical means .
Chaikin subdivision 2026-10-14
Chaikin subdivision generates a rounded subdivision curve by repeatedly cutting the corners off a control polygon. Each edge contributes two new control points, one a quarter of the way along the edge and the other three quarters of the way along it. Connecting these new control points in edge order replaces the old corners by shorter line segments. This makes Chaikin subdivision a useful way to draw a smooth outline from a small, editable set of control points, without evaluating B-spline basis functions at each drawing step. Joy presents both the corner-cutting rule and its closed version.[ref]
https://www.cs.unc.edu/~dm/UNC/COMP258/LECTURES/Chaikins-Algorithm.pdf. Kenneth I. Joy, Chaikin's Algorithms for Curves, §§Chaikin's Method and Example — A Closed Curve, UC Davis authored notes
For a closed control polygon with control points , set and interpret indices cyclically. One refinement givesand the new control polygon is the ordered sequenceClose it by joining back to . Use the old control points for every average in the same refinement; updating them in place would define a different algorithm. One refinement produces stored control points, and refinements produce , counting entries even if some coordinates coincide. Open control polygons need an explicit endpoint convention: applying only the interior edge rule discards the original endpoints, whereas retaining endpoints changes the boundary rule. The closed analysis below uses cyclic indices throughout.
For a worked example, start with the rectangular control polygonThe first edge gives and . Applying Chaikin subdivision to all four edges, including the closing edge, gives exactly eight control points in this order:Thus the rectangle becomes an octagonal control polygon. Two refinements have control points, and five have . The illustration shows finite control polygons at stages ; the last is already visually rounded but still consists of line segments.
The convex hull property follows directly from the weights. Each new control point is a convex combination of the endpoints of one old edge, because the weights are nonnegative and sum to one. Consequently, if is the convex hull of the stage- control points, then . Every later line segment, and every limit point, stays in the original convex hull. This gives a useful geometric enclosure. It also explains why Chaikin subdivision is approximating rather than an interpolation rule: the rectangle's four original corners are outside the first octagon's convex hull, so the limiting curve cannot pass through them. A refinement control polygon generally lies off the limiting curve as well.
For the cyclic uniform rule, the closed limit is the periodic quadratic uniform B-spline curve determined by the original control points. Uniform refers to equal spacing in the spline knot sequence, not equal physical lengths of edges of the control polygon. An infinite uniform control sequence has the corresponding local description; arbitrary open endpoint conventions need separate treatment. Joy derives the quarter-point refinement by splitting quadratic uniform B-spline spans.[ref]
https://www.cs.unc.edu/~dm/UNC/COMP258/LECTURES/Quadratic-B-Spline-Curve-Refinement.pdf. Kenneth I. Joy, Quadratic Uniform B-Spline Curve Refinement, §§Splitting and Refinement and The General Refinement Procedure, UC Davis authored notes
Here is an explicit description of that closed limit, with cyclic indices and a unit parameter span for each original control point:Each coordinate is a quadratic polynomial, and the three coefficients form a nonnegative partition of unity. The span runs from to . Differentiating gives and , so adjacent spans have matching values and matching first derivatives. Their second derivatives generally differ. Thus the parameterized curve is , generally not ; degenerate control points can still make its tangent vector vanish. For the rectangle, the span around iswhich rounds the original corner between edge midpoints and . For equally spaced abscissae, the centered Chaikin basis function gives the scalar graph version of this quadratic B-spline representation and has support of width three original grid intervals.
Another useful property is affine equivariance of a geometric basis: refining and applying an affine map commute. For ,and the same calculation holds for . The sum-to-one weights account for the translation. Therefore at every refinement, and also for the limit. A designer can apply a shear mapping, translation, or other affine map to the original control polygon and obtain exactly the transformed outline.
For example, compose the shear mapping with the translation to obtain . The rectangle's control points become . On its right edge, cutting first gives and then . Transforming first and cutting instead gives , the same control point.
Shearing and translating a control polygon commutes with one Chaikin refinement
. Left: the rectangle and its first refinement. Right: the transformed control polygon, with blue dots from cutting after the affine map and orange rings from applying the affine map after cutting. All eight pairs coincide. Details.Historically, George M. Chaikin published “An algorithm for high-speed curve generation” in Computer Graphics and Image Processing 3(4), December 1974, pages 346–349. Its emphasis was practical curve generation through a recursive algorithm using simple arithmetic, with applications including machine control and three-dimensional curves. The lasting appeal of Chaikin subdivision is how an elementary local average creates an entire smooth B-spline curve.[ref]
https://doi.org/10.1016/0146-664X(74)90028-8. G. M. Chaikin, An algorithm for high-speed curve generation, Computer Graphics and Image Processing 3(4), 346–349, 1974; publisher metadata and abstract
Closest point on a plane 2026-10-18
For a nonzero normal vector and the plane equation , the closest point on a plane to isIt subtracts only the normal displacement. Substituting gives , so lies on the plane; the difference is a scalar multiple of its normal vector. These two checks identify the orthogonal projection onto a plane without solving a three-variable minimization problem.
For , , and , the residual givesIndeed, , and . Choosing the plane point instead gives distance : being on the plane does not make a point the nearest one.
The closest point on a plane is reached along its normal vector
. Original three-dimensional example , with and . The finite mesh is only a displayed patch of the infinite plane; it is not a physical measurement. Details.As a function of , this projection isIt is an affine map, becoming a linear map when . Projecting a second time changes nothing. It preserves displacements along the plane and removes the normal displacement, so it is a concrete example of an idempotent map.
Common affine calibration cancellation in the lever rule is the exact invariance of the binary mixture phase fraction under the same affine map applied to the overall composition and both endpoints. For ,Both a shared offset and a shared scale cancel. The map must be common to all three measurements; separately fitted calibration curves or sample-dependent effects do not satisfy that assumption.
An original example has , , and . Applying gives measured values , and , still yielding . If only the overall-composition reading then receives an additional error, its inferred fraction changes by . Common calibration cancellation leaves independent sample errors intact.
The covariance matrix formulation explains why treating the shared errors as independent gives a false uncertainty. In input order , the Jacobian matrix row isIt annihilates both the common-offset direction and the common-scale direction , since . A joint covariance matrix representing only these shared affine influences therefore gives . The measurement guide explicitly identifies common instruments and reference data as sources of correlation that must be retained.[ref]
Guide to the expression of uncertainty in measurement by JCGM. §5.2.2, equation13; §5.2.4, common influences
Analyzing the overall sample and separated phases under one unchanged instrument calibration can realize this common contribution; repeating only one sample under another calibration breaks it. The cancellation is a property of this binary ratio. It does not remove nonlinear response, composition closure errors, endpoint misidentification or uncertainty in whether the samples establish liquid-liquid equilibrium.
Common-origin pressure-tap position error 2026-11-05
A common-origin pressure-tap position error shifts several inferred tap positions together relative to a fixed physical reference. It is a shared contribution to correlated position errors of pressure taps, not an independent placement error for every hole. Whether it affects a pressure-derived load depends on which references are held fixed.
For dimensionless chord coordinate , suppose the uncertain displacement is the same scalar for every tap. A first-order Taylor expansion of the pressure coefficient gives . For a reconstruction with fixed quadrature weights,The position covariance matrix is , a rank-one matrix. Squaring each tap's contribution before adding would discard the shared covariances. The covariance terms belong in the general propagation of uncertainty.[ref]
An original locally linear example takes eight equal weights , equal local slopes and . The reconstructed coefficient has standard uncertainty . Incorrect independence would give . If a two-dimensional load per span is with stipulated dynamic pressure Pa and airfoil chord m, these become and N/m. The chosen slopes describe a sensitivity model, not measured airfoil data.
Opposite-signed sensitivities can cancel the shared contribution. Conversely, repeating coordinate readings without independently checking the misplaced datum cannot average away its fixed error. A proposed test measures an independent reference mark and deliberately translates the assumed origin in the reconstruction, comparing the observed load sensitivity with the derivative above. If moving coordinates also changes the integration weights, those additional derivatives must be included.
Pressure-tap datum cancellation is a different situation. When the tap, leading edge and trailing edge are all expressed through the same affine map , with ,The common origin and scale cancel exactly. Merely translating a shared coordinate system causes no physical tap misregistration. Affine pressure-tap coordinate calibration must therefore state whether the datum is shared or independently fixed before assigning a common-origin uncertainty.
Composition of affine transformations 2026-10-31
If and , their function composition is , again an affine map. With column-vector homogeneous coordinates, its matrix is the product of the matrix for with that for , so the rightmost factor acts first.
Constant function Created 2026-09-24 Updated 2026-10-14
On a nonempty domain of a function , a constant function has one fixed value in its codomain:For example, on the real line is a horizontal graph. Its image of a function is the singleton even though its domain of a function has infinitely many inputs. A constant function from the real line to the real numbers has derivative zero and every nonzero real period of a function; there is no smallest positive period. It supplies a reference level from which variation can be measured.
A constant map between vector spaces is a linear map only when its value is the zero vector, since a linear map must preserve zero. A nonzero constant map is still an affine map. In a Fourier series, the constant term records the period average, while the other terms describe oscillation around that level.
Control-point influence on a curve 2026-10-14
For a parametric curve represented by fixed scalar functions and control points in Euclidean space, writeThe influence function of this handle is : the resulting displacement vector is , and its norm is . These are exact finite changes, not only a first-order estimate. Where is negative the displacement points opposite to . Where , the curve is unchanged at that parameter.
The influence coefficient 3t(1-t)^2 equals the vertical displacement for a unit upward move of P1, reaches 4/9 at t=1/3, and vanishes at both endpoints
. Details. For the cubic Bézier curve above, the Bernstein basis influence function is . Its derivative isIt increases up to and decreases thereafter on . Consequently the largest displacement at corresponding parameters isFor it is . This is a comparison at the same parameter. A geometric closest-distance comparison, as in a closest-point search on a subdivision curve, instead minimizes over another parameter:The closest parameter can differ from , so is not automatically a maximum closest-distance separation of the two curves. Neither parameter is assumed to be arc length.
The support of a function is the closure of the parameter values where . It bounds where a nonzero handle move can act, but may include boundary points where the actual displacement is zero. Here the support of a function is all of relative to the parameter domain, while the endpoints are unaffected. In contrast, a B-spline of order has support within in its spline knot sequence, so its handle affects only that parameter range; the support and Gram bandwidth of B-splines describes the same locality in another application. Changing the spline knots or the basis along with the handle requires a different comparison.
When the basis functions form a partition of unity, , each evaluation is an affine combination of the control points. Applying an affine map to all handles givesThis is affine equivariance of a geometric basis, also called affine covariance: affine maps preserve affine combinations. In particular, translating every handle by translates the whole curve by , whereas moving just one handle produces . The Bernstein basis has this partition of unity property.[ref]
Curves and Surfaces in Geometric Modeling: Theory and Algorithms by Jean Gallier. §4.3, pp. 132–133: Bernstein weights sum to one
Nonnegative weights further make evaluations convex combinations, giving the convex hull bound illustrated by the control polygon; nonnegativity is not needed for affine covariance.For a proof outline of the influence formula, subtract the two basis representations: every unchanged coefficient cancels, leaving . The same subtraction with several moved handles gives . No claim about interpolation, closest-point search on a subdivision curve or endpoint motion follows until the actual basis values are examined.
Danilo Jimenez Rezende 2026-11-05
Danilo Jimenez Rezende coauthored the 2014 stochastic-backpropagation work with Shakir Mohamed and Daan Wierstra, developing trainable generative and approximate-inference models.[ref]
Stochastic Backpropagation and Approximate Inference in Deep Generative Models by Danilo Jimenez Rezende and others. Primary full byline; §§2–4
With Shakir Mohamed, he subsequently developed normalizing flows for richer approximations in variational inference.[ref]Variational Inference with Normalizing Flows by Danilo Jimenez Rezende and Shakir Mohamed. Primary PDF full byline; §3, change-of-variables construction
An original one-dimensional transformation illustrates the density accounting. Let have standard normal distribution and take . Sampling is straightforward, but its density must also account for the stretching of length:At , this is . The factor comes from the absolute derivative of the inverse affine map. Omitting it would give a density integrating to instead of .
The example is only an affine Gaussian transformation; it does not claim that every nonlinear transformation preserves a Gaussian family. A normalizing flow composes suitable invertible transformations while tracking the corresponding density changes. This connects an efficient sampling procedure to the density evaluation needed in a variational objective, rather than treating the generated samples alone as a complete inference method.
De Casteljau's algorithm 2026-10-14
De Casteljau's algorithm evaluates a Bézier curve by repeated linear interpolation between neighboring control points. Fix , start with , and computeEach new row has one fewer point. The final point is . Use the same at every stage: it describes the fraction of each line segment traversed from its first endpoint toward its second. A cubic Bézier curve needs three levels containing , , and linear interpolations, respectively.[ref]
https://graphics.stanford.edu/courses/cs164-10-spring/Handouts/handout19.pdf. Stanford CS348a, Handout 19, §2.3, pp. 10–11
For the Bézier curve with control points , , , , evaluation at takes the midpoint of every neighboring pair. The first level givesThe second level givesand the third level givesThe intermediate control polygons shrink toward the evaluated point. Each step is a convex combination, giving a geometric explanation of the convex hull property; every step also commutes with an affine map.
Cubic Bézier curve, its convex hull and bounding box, and the three levels of De Casteljau's algorithm at t = 1/2
. Details. The same construction splits the Bézier curve exactly, with no approximation. The outer edges of the triangular array give the new control points: for a cubic, specify the left piece and specify the right piece.[ref]
https://graphics.stanford.edu/courses/cs164-10-spring/Handouts/handout19.pdf. Stanford CS348a, Handout 19, §2.3, p. 11, subsegment control points
At these two lists areWrite and for the Bézier curves defined by these lists, each with its own parameter . ThenTheir images reproduce the original curve, while their derivatives account for the shorter parameter intervals. At the join,The chain rule gives ; in the original parameter, both pieces have derivative . This is the interval scaling in the parametric first-derivative join of Bézier curves.Exact midpoint subdivision with two smaller control polygons, endpoint tangent directions, and equally spaced parameter samples
. Details. Repeated midpoint splitting is useful for drawing vector graphics paths: each piece has a smaller control polygon and its own convex hull enclosure. A renderer can refine pieces until they are sufficiently close to line segments at the chosen display scale. Evaluating and splitting it are thus two results of the same short algorithm.
For a proof outline, expanding the recursion givesPascal's identity proves this by induction; at , it is precisely the Bernstein basis formula for the Bézier curve. The geometric construction remains useful even when expanding the polynomial by hand would be cumbersome.
// Target: analysis.bigb









