Augmented matrix 2026-10-14
The augmented matrix of a linear system appends the right-hand side as one extra column. Elementary row operations preserve its solution set. The linear system is consistent exactly when the coefficient matrix rank equals the augmented matrix rank, because this is equivalent to lying in the span of the columns of . A larger augmented matrix rank produces a zero coefficient row with a nonzero right-hand side during Gaussian elimination.
A complex nilpotent matrix of size six is determined up to matrix similarity by its matrix rank and nilpotency index. Its Nilpotent Jordan block sizes form an integer partition of six. If there are blocks and the largest size is , then its matrix rank is and its minimal polynomial is . The eleven possible size lists areNo two have the same pair , which proves the assertion by Jordan normal form. The dimension restriction matters: in size seven, block lists and both have matrix rank four and minimal polynomial , but give different Jordan normal forms.
Column space Created 2026-09-29 Updated 2026-10-18
The column space of a matrix is the linear span of its columns. It is the image of a linear map represented by .
If an matrix has columns , then matrix multiplication gives . Thus its column space consists of exactly the attainable output vectors in , and has a solution exactly when belongs to this column space. Its dimension is the matrix rank; the kernel instead describes input vectors in whose contributions cancel.[ref]
Solution Sets. Interactive Linear Algebra, §2.4.3, Solution Sets and Column Spans: attainable outputs and input solution sets occupy different spaces
Computational topology 2026-11-05
Computational topology develops algorithms for topological questions about finite representations of spaces, such as simplicial complexes. It connects discrete data structures to invariants that ignore irrelevant geometric distortion. connected components and homology groups are examples of computable structural information.[ref]
For a small simplicial homology calculation over a field, take three vertices joined by three edges in a cycle, without a filled face. The matrix of the first boundary operator has matrix rank two, so this linear map has a kernel of a linear map of dimension one; with no two-dimensional face, that cycle is not a boundary. Thus the first Betti number is one. Adding the filled triangular face makes the same cycle a boundary and reduces the first Betti number to zero. The vertex coordinates need not change: the encoded incidence structure determines whether the hole is present.
Critical point of a smooth map 2026-10-14
Imagine folding a sheet along its horizontal axis by the smooth map , . The inputs and have the same output. Along , horizontal motion still changes the output, but vertical motion produces no first-order change: . These inputs are critical points of a smooth map even though their differential is not the zero linear map. The missing direction in the target tangent space is what matters.
Precisely, let be a smooth map between smooth manifolds without boundary. An input is a critical point of a smooth map when its differentialis not surjective, equivalently when its rank of a linear map is less than . An input where it is surjective is a regular point; a map with no critical points of a smooth map is a submersion. The threshold is the target dimension, not . In manifold charts, compute the matrix rank of the Jacobian matrix; changes of charts preserve that matrix rank. [ref]
For the fold,Thus its set of critical points of a smooth map is the entire horizontal axis. Away from that axis, its differential can produce every infinitesimal target displacement. At the axis, it can produce only horizontal ones. This diagnoses where the two inverse branches in the picture merge.
For a scalar smooth function , the target dimension is one. A linear map to fails to be surjective exactly when it is zero, so here a critical point of a smooth map means . In Euclidean space this is precisely the familiar critical point condition : every first partial derivative vanishes. On a Riemannian manifold, it is equivalently for the chosen Riemannian metric; the Riemannian gradient depends on that metric, but whether it vanishes does not. See Loring W. Tu, An Introduction to Manifolds, Proposition 8.23 for the scalar coordinate criterion. The scalar smooth function has a critical point at zero without a local maximum or local minimum.
A source of lower dimension gives a different warning. For the smooth map , , the Jacobian matrix is the column , of matrix rank one at every input. Every input is a critical point of a smooth map, since the target dimension is two, even though this parametric curve has nonzero velocity everywhere and is an immersion. More generally, if , no differential can be surjective, so every input is critical. The terminology measures the ability to fill target directions, rather than whether the parametrized curve looks smooth.
Critical value 2026-10-14
A critical point of a smooth map is an input; a critical value is an output reached from at least one such input. For the fold , all inputs are critical points of a smooth map, and all outputs are critical values. The identical coordinate pairs live in different copies of . To classify an output , inspect its entire fiber of a function:For both inputs have matrix rank two, so the output is a regular value. For its input has matrix rank one, so the output is a critical value. For the fiber of a function is the empty set, and the output is a regular value despite being outside the image of a function.
In general, for a smooth map , write for its set of critical points of a smooth map. Its set of critical values is . Equivalently, is a critical value if some has a differential that is not surjective. A regular value requires a surjective differential at every input in that fiber of a function. The universal condition holds vacuously for an empty fiber of a function: every output outside is a regular value. These quantifiers are explicit in Loring W. Tu, An Introduction to Manifolds, Definition 8.22 and its two following remarks.
Regular circular fibers and a critical value reached from both critical and regular inputs
. Details. The distinction is useful when solving equations. For a smooth map between smooth manifolds without boundary and a regular value with nonempty fiber of a function, the regular level set theorem makes an embedded submanifold of dimension , with tangent spaceThe implicit function theorem explains the freedom count: locally, independent equations fix coordinates and leave free. The rank-nullity theorem gives the same count for the displayed kernel of a linear map. See Eckhard Meinrenken, Differentiable Manifolds, Theorems 1.11 and 1.15 for the local projection description and its consequence for level sets.
For example, , , has gradient and just one critical point, the origin, with critical value zero. For , its fiber of a function is the circle of radius , a one-dimensional embedded submanifold; negative are regular values with empty fibers. At , the fiber of a function is a single point instead of a circle. A critical value does not force its fiber of a function to be nonsmooth: that singleton is itself a zero-dimensional embedded submanifold, but the regular dimension conclusion does not apply.
Nor does a single regular input certify a regular value. The scalar smooth function has and derivative . The input zero is a critical point, while the input one is regular because . Their shared output zero is nevertheless a critical value. Conversely, different critical points of a smooth map can share one critical value: sends both critical inputs and to zero. Counting inputs and counting critical outputs are different tasks.
Finally, for , every output on the parabola is a critical value, and every output off it is a regular value with empty fiber of a function. There is no nonempty regular fiber of a function here, so the negative expression never becomes a claimed manifold dimension. Both diagrams show exact mathematical examples, not measured data.
Derek Jones (magnetic resonance researcher) 2026-11-05
Derek Jones is a physicist at Cardiff University whose research develops quantitative magnetic resonance imaging, especially diffusion tensor imaging. He trained in physics at Nottingham, medical physics at Leeds and completed a Leicester doctorate on diffusion tensor imaging. His work asks how acquisition design changes what can be inferred from a measured signal.[ref]
Derek Jones profile at Cardiff University. Overview, Research and Biography: research focus and1993/1995/1998 education
Derek Jones, from his labeled Cardiff University staff-profile portrait
. Jones investigated how diffusion MRI gradient sampling affects the measurement uncertainty of inferred diffusion tensors. In 2004, Jones used Monte Carlo simulations to compare sampling schemes at a diffusion weighting b-value of , tracking tensor orientation, trace and fractional anisotropy. It showed why a minimal set of directions need not give robust estimates; its numerical findings belong to the tested models and noise conditions.[ref]
The effect of gradient sampling schemes on measures derived from diffusion tensor MRI: A Monte Carlo study by Derek K. Jones. Original abstract: controlled b-value, tensor-derived outcomes and comparison of angular sampling schemes
An earlier conference report made the comparison with a fixed observation budget: one baseline and 60 diffusion-weighted observations. Six unique orientations repeated 10 times and 30 orientations repeated twice therefore used the same number of diffusion-weighted observations. Varying angular coverage while controlling that budget separates a direction-design question from simply acquiring more data.[ref]
The Effect of Gradient Sampling Scheme on Estimates of Fiber Orientation: Implications for Fiber Tractography by D. K. Jones. Methods and Table1:61 total rows, unique orientations and repetition counts
A noise-free calculation illustrates diffusion tensor identifiability. With separately known baseline , fixed and a unit direction , the Gaussian diffusion tensor model givesThe symmetric matrix has six independent coefficients. Its quadratic form produces a row in a linear least squares design. Directions give six observations but only matrix rank three: opposite directions repeat a row and the off-diagonal coefficients remain undetermined.
Adding the normalized mixed directions , and to the three positive coordinate axes instead gives matrix rank six. For example,This establishes uniqueness in the stated noise-free model. It does not establish small measurement uncertainty or an optimal experimental design.
Original noise-free diffusion-tensor design matrices: six antipodal directions give rank three, while axes plus three mixed directions give rank six
. Details. Jones also interviewed John E. Tanner about the physical origins of pulsed field gradient NMR. The interview preserved concrete details of the apparatus, including Tanner's carefully spaced mica stack and the practical difficulties of rapidly switching magnetic field gradients.[ref]
John Tanner, the return to Germelshausen, and the work that nearly never happened by Derek Jones. pp.6–7: restricted-diffusion apparatus and switching problems
Diffusion MRI gradient sampling 2026-11-05
Diffusion MRI gradient sampling is the choice of magnetic field gradient directions, strengths and repetitions used to acquire diffusion magnetic resonance imaging measurements. Under a Gaussian diffusion tensor model, different directions probe different values of a quadratic form. Acquiring many measurements in an uninformative set of directions cannot determine every tensor coefficient.
For a unit gradient direction , a separately measured matched baseline and fixed positive diffusion weighting b-value , the ideal signal model isThe symmetric matrix has six independent coefficients. Opposite directions and give the same model prediction. Thus counting orientations without considering their independent information can overstate diffusion tensor identifiability.
A 2003 Monte Carlo simulation study controlled the observation budget: one baseline plus60 diffusion-weighted observations. Six distinct orientations repeated ten times and30 distinct orientations repeated twice each use60 diffusion-weighted observations. The comparison asks how angular coverage and repetition affect estimated orientation, rather than simply comparing a longer acquisition with a shorter one.[ref]
The Effect of Gradient Sampling Scheme on Estimates of Fiber Orientation: Implications for Fiber Tractography by D. K. Jones. Methods and Table1: observation budget, directions and repetitions
The subsequent 2004 study compared direction schemes through estimated orientation, trace and fractional anisotropy at . These are distinct outputs of the same tensor fit. An acquisition suited to one readout need not minimize uncertainty in another.[ref]
The effect of gradient sampling schemes on measures derived from diffusion tensor MRI: A Monte Carlo study by Derek K. Jones. Primary Abstract: Monte Carlo design and tensor-derived measurements
An original design comparison separates two questions. Repeating the three coordinate axes may reduce random variation in their measured attenuations, but it adds no equations for the three off-diagonal tensor entries. Adding suitable mixed directions can supply those equations. Full matrix rank establishes uniqueness in the ideal model; practical measurement uncertainty, motion and model mismatch still require assessment.
Diffusion tensor identifiability 2026-11-05
Diffusion tensor identifiability asks whether the measurements determine a unique diffusion tensor within a stated model. For a constant symmetric matrix , known matched baseline and known , defineFor unit directions , these equations give a linear system with rowAll six coefficients are uniquely determined in this unrestricted noise-free linear model exactly when . The condition concerns independent equations, not merely the number of acquired images. Direction design and uncertainty in tensor-derived measurements were examined experimentally through controlled simulations by Derek Jones.[ref]
The effect of gradient sampling schemes on measures derived from diffusion tensor MRI: A Monte Carlo study by Derek K. Jones. Primary Abstract: sampling comparisons; displayed rank criterion is an explicit linear-algebra derivation
For a counterexample, use the six directions . Each antipodal pair repeats a row, so the matrix rank is three. Only are determined. Even restricting to physically positive tensors does not generally repair this: the identity matrix and a matrix with unit diagonal and both fit those data for any , while having different off-diagonal entries.
Replace the negative axes by , and . The axes first give the diagonal entries, thenand similarly for and . This directly proves uniqueness for the new design. If , and , all in units of , then in those units.
Original design-matrix example: antipodal axes give rank three, while axes and mixed directions give rank six
. Reused algebra illustration, not a reconstruction of a published acquisition or simulation. Rows represent measurements; columns represent the six independent tensor coefficients. Details.Full matrix rank is not a guarantee of a stable estimate in noisy data. If must also be fitted, its logarithm is another unknown; the augmented signal model needs a corresponding independent baseline constraint. The six-coefficient argument above explicitly assumes is separately known. Likewise, tractography requires additional interpretation beyond uniqueness of a local tensor.
Eckart-Young-Mirsky theorem Created 2026-10-14 Updated 2026-10-25
For a finite matrix with singular values in nonincreasing order, truncating its singular value decomposition after terms minimizes the Frobenius norm error, whose square is the sum of the discarded squared singular values. It also minimizes the operator norm error, which equals the next singular value, or zero if none remain. See Mazeika, Section 4.3: Low rank approximation.
The problem of replacing a matrix by one of lower matrix rank is explicit in Eckart and Young's 1936 paper, [ref]
https://www.cambridge.org/core/journals/psychometrika/article/approximation-of-one-matrix-by-another-of-lower-rank/B29672E1EDD0FA1B7611D4DFAFC321B3. The approximation of one matrix by another of lower rank, Psychometrika 1, 211-218
. The publication predates the modern image compression example, showing how a general matrix question can support later applications. Effective rank of a covariance matrix Created 2026-09-28 Updated 2026-11-05
For a nonzero positive semidefinite covariance matrix , its covariance effective rank iswhere the operator norm is induced by Euclidean norm and the are the nonnegative eigenvalues. Each nonzero eigenvalue contributes at most one to this ratio, soA scalar multiple of has covariance effective rank , while a matrix with and a unit vector has covariance effective rank one. For , the ratio is , although the ordinary matrix rank is three.
Thus the quantity measures total variance relative to the largest directional variance, rather than merely counting nonzero directions. Multiplying by a positive scalar leaves the ratio unchanged. Its role in the Gaussian sample-covariance operator-norm bound makes it a useful measure of the dimension that matters for covariance estimation. The zero matrix requires a separate convention because the displayed ratio is .
Elementary matrix Created 2026-10-14 Updated 2026-10-25
An elementary matrix is obtained by applying one elementary row operation to the identity matrix. Multiplying by it on the left performs the same row operation on . Such a matrix is invertible: reverse a row swap, reverse a nonzero row scaling by its reciprocal, or reverse adding a multiple of another row by adding its negative. Right multiplication gives an elementary column operation. In particular these operations preserve matrix rank.
Equicorrelation covariance matrix Created 2026-10-14 Updated 2026-11-05
For variables with unit variances and common pairwise correlation , the equicorrelation covariance matrix isOn the constant-vector direction, has eigenvalue ; on its orthogonal complement, it has eigenvalue zero. Hence the eigenvalues of are and , the latter with multiplicity . Positive semidefiniteness is therefore equivalent toAt , the matrix has matrix rank one. At , the constant direction has zero variance and the matrix has matrix rank . In the interior it is positive-definite. For , the first population principal component is the normalized sum of the centered variables; for , the largest eigenvalue instead occurs on the orthogonal complement.
The dimension restriction matters. For , the formula gives for every , so the lower bound involving does not apply: there is no distinct variable pair whose correlation could determine .
Finite-dimensional quantum compression converse Created 2026-10-06 Updated 2026-11-05
If encoding and decoding quantum channels factor through a -dimensional space, their composite Kraus operators all have matrix rank at most . The rank bound for a weighted operator trace and the Kraus formula for entanglement fidelity implyThus high operation fidelity requires the leading eigenvalues of the source density operator to carry almost all its probability. The bound applies to arbitrary completely positive trace-preserving encoders and decoders.
Every real symmetric matrix satisfying for every real vector is the covariance matrix of a random vector taking only finitely many values. The stated inequality means that is a positive semidefinite matrix. One does not need a multivariate normal distribution to realize a specified second-moment geometry. Conversely, a proposed covariance matrix must be positive semidefinite, since is the variance of the projection onto any real vector .
Here is an explicit construction. If is the matrix rank of , orthogonal diagonalization gives a real matrix withLet be independent Rademacher random variables, each equal to or with probability , and put . Then and , soThe random vector has at most possible values. Adding any constant vector prescribes its mean without changing its covariance matrix. If , the constant vector alone suffices.
For an original two-dimensional example chooseThe four equally likely values of areTheir coordinate means vanish. Their squared first coordinates average to , their squared second coordinates to , and their coordinate products to . The displayed covariance matrix is therefore realized exactly, without sampling error or an asymptotic argument.
These four values are a complete probability model with masses , not automatically an independent random sample whose population is unknown. Applying the usual divisor-three sample covariance matrix to the same four rows would give . NIST explicitly distinguishes that unbiased sample formula with divisor ; probability-weighted population moments use the assigned masses instead.[ref]
NIST and SEMATECH e-Handbook of Statistical Methods. Section6.5.4.1, Mean Vector and Covariance Matrix, unbiased sample formula
The sign construction need not minimize the number of points. If a distribution has positive-mass support points, the centered points obey one nontrivial weighted relation, so their span has dimension at most . Its covariance matrix consequently has matrix rank at most , requiring when . Matching a covariance matrix also does not match the full probability distribution: this finite model and a continuous Gaussian model can have identical means and covariances while assigning very different probabilities to nonlinear events.
Full column rank Created 2026-09-28 Updated 2026-10-18
An matrix has full column rank when its matrix rank is , equivalently when its columns are linearly independent vectors. This requires . Its kernel is , so the represented linear map is injective and every solvable linear system has a unique solution. It need not reach every output.[ref]
One-to-one and Onto Transformations. Interactive Linear Algebra, §3.2.1, One-to-one matrix transformations theorem and Wide matrices do not have one-to-one transformations
For example,has full column rank two, but its column space is the plane in . An attainable output fixes the input uniquely as ; an output outside this plane has no preimage.
Full row rank Created 2026-09-28 Updated 2026-10-18
An matrix has full row rank when its matrix rank is , equivalently when its rows are linearly independent vectors. This requires . Its column space is all of , so the represented linear map is surjective and every linear system is solvable. If , its nullity is , so solutions are not unique.[ref]
One-to-one and Onto Transformations. Interactive Linear Algebra, §3.2.2, Onto matrix transformations theorem and Tall matrices do not have onto transformations
For example, the matrix transpose of the matrix in full column rank isIt has full row rank two. Every output has solutions , with . A square matrix of size has full row rank and full column rank together exactly when its matrix rank is ; then it is an invertible matrix. A nonsquare matrix cannot have both properties, and has no two-sided inverse: such an inverse would make its linear map a linear isomorphism between vector spaces of different dimensions.
Full-row-rank matrix manifold 2026-09-28
The real by matrices of matrix rank form an open smooth manifold . Indeed, it is the union of the open sets on which a chosen by minor has nonzero determinant, so its dimension is .
Fundamental matrix (computer vision) 2026-11-05
The fundamental matrix encodes the epipolar geometry of two distinct camera centers using image homogeneous coordinates. For matching image points , the ideal constraint isThus is an epipolar line in the right image, while is an epipolar line in the left image. A nondegenerate fundamental matrix has matrix rank two and is defined only up to nonzero scale.[ref]
Epipolar Geometry in the OpenCV-Python tutorials. Basic Concepts: correspondence constraint, rank-two matrix and point-to-line mapping
For an ideal horizontally aligned stereo pair with matched image coordinate origins, one possible scale convention isPutting gives , so the right point must satisfy . Its horizontal coordinate remains free: the constraint reduces the search to a line but does not determine stereo disparity or depth. Repeated texture can still make stereo correspondence ambiguous.
With camera calibration matrices and relative pose under the convention , the relation is , where . Translation direction affects the geometry, but multiplying by a scalar only rescales : the fundamental matrix alone cannot establish a metric camera baseline. Pure rotation has and requires a different geometric description.[ref]
Epipolar Geometry in the OpenCV-Python tutorials. Basic Concepts: essential and fundamental matrices, intrinsic calibration and degrees of freedom
Gaussian error Created 2026-10-31 Updated 2026-11-05
A Gaussian error is a random error term with a Normal distribution. The exact least-squares conclusions require a specified model: for fixed-design linear regression, letwhere the design matrix has full column rank. The error vector thus has independent zero-mean Gaussian coordinates of common variance. For fixed , maximizing the likelihood over is equivalent to minimizing .
The least-squares estimator and its exact sampling law areThis follows because a linear transformation of a Gaussian vector is Gaussian. It is an exact finite-sample law, rather than merely Asymptotic normality.
Zero mean and common variance alone do not imply Gaussian errors or exact normal coefficient estimates. Even with Gaussian errors, fitted residuals are not independent copies of them: with , the residual vector is and has covariance , usually with correlated coordinates and reduced rank. A Normality assumption must identify whether it concerns errors, responses or another latent variable.


