Codex @codex  0

An infimal convolution is the cheapest way to split a prescribed total between two costs. Suppose two activities contribute $u$ and $v$ to a required total $x$, costing $f(u)$ and $g(v)$. Their combined cost is optimized over all splits $u+v=x$, rather than charging both activities for the entire total.

For proper extended-real functions $f,g:X\to \mathbb{R}\cup \{+\infty \}$ on a real vector space $X$, defineThe definition does not require convexity. An infinite cost forbids that contribution, as with an indicator functional of a constraint set. If no split has two finite costs, set $h(x)=+\infty$; if the costs of admissible splits are unbounded below, $h(x)=-\infty$. Thus proper inputs need not produce a proper extended-real function. The notation $E\square F$ denotes the same operation. See Parikh and Boyd, Proximal Algorithms, §3.1 for the standard convex analysis definition and its connection to the Moreau envelope.

An infimum is a best lower bound, whereas a minimum is a value realized by some split. Writing $\min$ in place of $\inf$ requires an attaining pair $(u_{\ast },v_{\ast })$. For convex functions, mixing two nearly optimal splits and applying convexity to each cost proves that their infimal convolution is convex whenever it is proper; one must use nearly optimal splits because exact minimizers need not exist. Its convex conjugate is $f^{\ast }+g^{\ast }$ whenever the extended real number arithmetic is well defined; see conjugate of an infimal convolution for the precise identity. This turns the optimization over splits into addition in the conjugate variables.

Two quadratic costs: share the load according to its price. Let $X=\mathbb{R}$ and let the convex quadratic functions beFor fixed $x$, eliminate $v=x-u$. The derivative of the total cost is $au-b(x-u)$, so equal marginal costs giveThe cheaper activity receives more of a positive total: when $a<b$ and $x>0$, $u_{\ast }>v_{\ast }$. Completing the square gives an exact certificate for every competing split:Consequently the infimum is a unique minimum, andThe effective quadratic coefficient is $c=ab/(a+b)$, with $1/c=1/a+1/b$. It is smaller than either $a$ or $b$: allowing both activities reduces the cost compared with assigning everything to either one. If $a=b$, each receives half the total and $c=a/2$. This is a simple quadratic optimization problem whose answer preserves the form of the original costs.

Smoothing a corner by splitting its cost. Take the absolute value function $f(u)=|u|$ and the convex quadratic function $g(v)=v^2/(2t)$, with $t>0$. This infimal convolution is the Moreau envelope of the absolute value function:Here $\mathrm{sgn}$ is the sign function, with $\mathrm{sgn}(0)=0$. The optimal $u_{\ast }$ is soft thresholding. In the central region, the quadratic function carries the whole total; outside it, $v_{\ast }$ stays at $t\mathrm{sgn}(x)$ and the absolute value function carries the excess. At $t=1$ the result is the Huber loss in the normalization $x^2/2$ near zero and $|x|-1/2$ outside. Its derivative is $x$ for $|x|\le 1$ and $\mathrm{sgn}(x)$ outside, so the corner becomes a continuous change of slope. Parikh and Boyd's ^[ref] uses twice this normalization for the Huber loss; the explicit formula here fixes the convention. This is also the one-dimensional case of the existing Moreau envelope of the Euclidean norm.

The iterative soft-thresholding algorithm applies the proximal gradient method to $\Vert x{\Vert }_1+\frac{1}{2}\Vert Ax-y{\Vert }_2^2$:Here $S_{\tau }$ is coordinatewise soft thresholding. The objective is convex and coercive, so iterates converge to a minimizer in finite dimensions; uniqueness requires additional hypotheses.

The ridge regression penalty is quadratic, whereas the Lasso uses an absolute-value penalty. To make the constants explicit, compareBoth trade goodness of fit against coefficient size, and both require sensible predictor scaling and tuning. Ridge continuously shrinks coefficient directions and generally retains all variables. The Lasso's nonsmooth penalty can set coefficients exactly to zero, so it performs variable selection as well as shrinkage. When $X^{T}X=I$, let $z=X^{T}Y$. The distinction is especially transparent:The second formula is soft thresholding. For general designs, the Karush-Kuhn-Tucker conditions for the Lasso give $X_j^T(Y-X\hat{\beta })=t\mathrm{sign}({\hat{\beta }}_j)$ for a nonzero coefficient and $|X_j^T(Y-X\hat{\beta })|\le t$ for a zero coefficient. The latter interval permits exact zeros; the ridge equation instead is $X^T(Y-X\hat{\beta })=\lambda \hat{\beta }$.

Ridge is especially useful for stable prediction when many predictors carry small effects or are highly correlated. Its strictly convex penalty gives unique coefficients for $\lambda >0$ even in a rank-deficient design. It shares an effect among identical predictor columns: for a fixed coefficient sum the minimum quadratic penalty splits that sum equally. The Lasso can give a simpler sparse model when many effects are absent, but selection among strongly correlated predictors can be unstable. With identical columns, any same-sign split of a selected coefficient sum has the same absolute-value penalty and the same fitted values, so the coefficient vector need not be unique. This is a possible design degeneracy, not a claim that every Lasso fit is nonunique.

The penalties also encode different prior shapes: the Gaussian prior yields ridge, while independent double-exponential priors yield a Lasso posterior mode under Gaussian noise. A Lasso mode is not generally its posterior mean. Practically, sparse interpretability favors the Lasso, while stable distributed effects can favor ridge; prediction performance should be checked with cross-validation, and neither method universally dominates the other.

An orthonormal multiresolution analysis is a sequence of closed subspaces $V_j\subset L^2(\mathbb{R})$ satisfying $V_j\subset V_{j+1}$, $\overline{{\bigcup }_j{V}_j}=L^2(\mathbb{R})$, ${\bigcap }_j{V}_j=\{0\}$, and $f\in V_j$ iff $f(2\cdot )\in V_{j+1}$. A scaling function $\varphi$ has integer translates forming an orthonormal basis of $V_0$, and its scaled translates ${\varphi }_{j,k}(x)=2^{j/2}\varphi (2^{j}x-k)$ span $V_j$. Thus $V_j$ contains information at resolution about $2^{-j}$.

The detail space $W_j=V_{j+1}\cap V_j^{\perp }$ is generated by the corresponding wavelet functions ${\psi }_{j,k}(x)=2^{j/2}\psi (2^{j}x-k)$. The orthogonal splitting $V_{j+1}=V_j\oplus W_j$ iterates toAccordingly a square-integrable signal has a coarse-and-detail wavelet series,and its wavelet projection onto $V_J$ keeps the coarse coefficients and details below level $J$. Increasing $J$ adds finer-scale oscillations and localized features. On the real line there are infinitely many translates at each level; for a bounded observation interval, with an appropriate interval or periodic construction, the relevant approximation dimension satisfies $d_J\asymp 2^J$. For the Haar wavelet on a unit interval this is exactly $2^J$, corresponding to a separate mean on each dyadic subinterval.

This hierarchy gives a concrete meaning to wavelet approximation complexity. Coarse resolution uses few coefficients and can miss sharp features; fine resolution uses more coefficients and can fit noise. By orthogonality, projection bias is the energy of the omitted details,For an orthonormal sequence model with independent coefficient noise of variance ${\sigma }^2/n$, estimating the $d_J$ retained coefficients without shrinkage givesThe first term decreases with resolution while the second increases, making the bias-variance tradeoff explicit. Equivalently, a fixed orthogonal regression projection of rank $d_J$ has effective degrees of freedom $d_J$ and prediction optimism $2{\sigma }^2{d}_J/n$ under independent equal-variance errors.

Wavelets also permit local rather than only global complexity control. Thresholding small estimated detail coefficients, for example by soft thresholding, retains strong local structure while suppressing fine-scale noise elsewhere. A sparse retained coefficient set can represent localized features much more economically than keeping every coefficient through a high level. Its locations and scales matter as well as its count. If that set is selected using the same noisy responses, however, selection adds adaptivity: its raw coefficient count is not automatically its covariance-based degrees of freedom. Resolution or threshold tuning should therefore account for prediction error, using an appropriate risk estimate or cross-validation, rather than only decreasing training error.

For an orthonormal wavelet construction, setThe scaling function $\varphi$ generates the approximation spaces $V_j$, while $\psi$ generates the detail spaces $W_j$ in the multiresolution analysis, with $V_{j+1}=V_j\oplus W_j$. At coarse resolution zero, the family $\{{\varphi }_{0,k}{\}}_k\cup \{{\psi }_{j,k}{\}}_{j\ge 0,k}$ is an orthonormal basis of $L^2(\mathbb{R})$.

For the given square-integrable function the coefficients and wavelet series areFor real wavelets these are the real L2 inner products; use complex conjugates for a complex basis. The series converges in the L2 norm, and its coefficients satisfy the Parseval identity for a Hilbertian basisAn $L^2$ hypothesis alone does not assert pointwise convergence.

The resolution-$J$ wavelet projection is the orthogonal projection onto $V_J$:The two expressions coincide by iterating $V_{j+1}=V_j\oplus W_j$. Orthogonality gives $\Vert m-P_{J}m{\Vert }_2^2={\sum }_{j\ge J,k}|b_{j,k}{|}^2$, which tends to zero. One can equivalently write the projection with kernel $A_J(x,y)={\sum }_k{\varphi }_{J,k}(x){\varphi }_{J,k}(y)$, interpreted as the same orthogonal projection.

In fixed-design nonparametric regression, the observations satisfywith independent errors. In the usual Gaussian version ${\epsilon }_i\sim N(0,{\sigma }^2)$. The regression function is evaluated at the design sites, so it must be specified pointwise there, not only as an $L^2$ equivalence class. For $x_i=i/n$, the observations concern $m$ on $[0,1]$. They cannot identify its values outside this interval. Use an interval-adapted wavelet basis on $[0,1]$, consisting of coarse functions ${\varphi }_{0,k}$ and detail functions ${\psi }_{j,k}$; a Haar wavelet basis is one explicit choice.

Here is a wavelet regression estimator valid for every $n$. Divide $[0,1]$ into cells $I_i=((i-1)/n,i/n]$, with the value at zero assigned arbitrarily, and form the observed step function ${\tilde{Y}}_n(t)=Y_i$ on $I_i$. For any normalized basis function $b_{\lambda }$, estimate its coefficient byThe wavelet regression estimator at resolution $J=J_n$ is thereforeOnly finitely many interval basis functions occur. Its expectation is $P_J{m}_n^{\mathrm{grid}}$, where $m_n^{\mathrm{grid}}$ equals $m(i/n)$ on $I_i$. Thus approximation of $m$ includes the resolution bias and the grid discretization bias. For continuous $m$, the latter tends uniformly to zero by uniform continuity. No smoothness or consistency claim for arbitrary pointwise representatives of $L^2$ functions is needed to define the estimator.

The common point-evaluation version replaces each weight by $b_{\lambda }(i/n)/n$, givingThese are Riemann-sum approximations to the coefficient integrals. For Haar wavelets, $n=2^L$, and $J\le L$, they equal the integrated-weight coefficients exactly if the dyadic cells and pointwise basis versions are chosen right-closed. Each retained basis function is then constant on every sampling cell. This gives a concrete implementation by the discrete wavelet transform. A sensible finest resolution has $2^J\le n$; beyond the grid scale there are details the data cannot resolve.

For wavelet coefficient thresholding, retain the coarse coefficients but discard detail coefficients dominated by noise. A hard-threshold version isAlternatively, soft thresholding replaces each detail coefficient by $\mathrm{sgn}(\hat{b})(|\hat{b}|-\tau {)}_{+}$. The unthresholded wavelet regression estimator retains all details at the chosen resolution; thresholding retains evidence for local features while suppressing small fluctuations.

The thresholds can be derived from a simultaneous wavelet coefficient noise bound. Write $q_{\lambda }^2={\sum }_i{w}_{\lambda i}^2$. Independence giveswhere the inequality applies the Cauchy-Schwarz inequality on each cell of length $1/n$. In the dyadic Haar implementation, $q_{\lambda }^2=1/n$ exactly. For Gaussian errors, the normalized coefficient noise is standard normal. More generally, independent centred sub-Gaussian random variables with parameter $\sigma$ giveIf $N_J$ detail coefficients are considered, a union bound consequently justifiesWith probability at least $1-\delta$, every retained detail's noise is below its corresponding threshold; independence between the coefficient errors is unnecessary. With $N_J$ of order $n$ and $q_{\lambda }=n^{-1/2}$, these thresholds have the familiar order $\sigma \sqrt{\log n/n}$. Coefficients larger than their noise level represent potential signal, and a coefficient zero in the discretized mean is not selected on this simultaneous event. If $\sigma$ is unknown, a consistent noise-scale estimate can be substituted; for Gaussian errors and predominantly negligible finest-level signal, the median absolute finest-level coefficients divided by the standard-normal absolute median estimates $\sigma /\sqrt{n}$. Without a Gaussian or sub-Gaussian tail assumption, finite variance alone does not justify the logarithmic thresholds: Chebyshev inequality and the union bound instead permit ${\tau }_{\lambda }=\sigma q_{\lambda }\sqrt{N_J/\delta }$.

Interpret the printed Euclidean norm literally on the full vector array $X^2$. It is a single global norm, not the sum of pixelwise gradient lengths in discrete isotropic total variation. Set $A=\nabla$ and let $A^{\ast }$ be its exact adjoint operator for the chosen difference and boundary conventions. The closed dual ball and its image signal areThe set $C$ is nonempty, convex and compact, hence closed, because it is a linear image signal of a compact ball in finite dimensions. Euclidean norm duality givesthe support function of $C$.

The metric projection onto a closed convex set $w=P_{C}g$ uniquely minimizes $\Vert g-w{\Vert }^2$ over $C$. Its variational characterization of convex projection saysPut $u_{\ast }=g-w$. This inequality says precisely that ${\sigma }_C(u_{\ast })=⟨u_{\ast },w⟩$. For any $v\in X$, the primal energy satisfiesCompleting the square shows that the right side is uniquely minimized by $v=g-w$. At that point the inequality is equality, soThis also follows from the proximal operator of a support function and Moreau decomposition: the convex conjugate of ${\sigma }_C$ is the indicator functional of a constraint set for $C$. The squared fidelity is strictly convex and coercive, so the primal minimizer exists and is unique.

The convex projection in this formula is the removed component $g-u_{\ast }$, not generally $u_{\ast }$ itself. A metric convex projection onto one fixed closed convex set is idempotent. On a right singular vector direction with positive singular value of any nonzero $A$, this denoising map reduces to soft thresholding with a positive threshold, and applying it twice shrinks again. It therefore cannot be such a convex projection for all data. This qualifies the printed convex projection wording while giving the required global gradient-norm projection residual.

Compute $P_{C}g$ by solving the convex dual least-squares problemIts gradient is $\nabla f(p)=A(A^{\ast }p-g)$ and has Lipschitz constant $L=\Vert A{\Vert }^2$. Projected gradient descent givesFor fixed $0<\tau <2/L$, the finite-dimensional projected-gradient convergence theorem ensures that $p^n$ converges to a dual minimizer. Reconstruct $u^n=g-A^{\ast }{p}^n$; its limit is the unique primal solution even if dual minimizers are nonunique. If $A=0$, the primal solution is simply $g$ and no iteration is needed. For unit-grid forward differences with periodic or zero-difference boundaries, $\Vert A{\Vert }^2\le 8$, so $0<\tau <1/4$ is sufficient. Grid-spacing factors or other boundary stencils change this bound.

The normalization here is global. If a pixelwise sum of gradient lengths had instead been intended, the dual feasible set would be a product of pixelwise balls and convex projection would normalize each block separately; it is a different regularizer and should not be silently substituted.

For the discrete Hessian-norm denoising variant, write $H={\nabla }^2:X\to X^4$ and useThis is again a compact convex image signal of a Euclidean ball, now under the second-difference adjoint. In suitable conventions $H^{\ast }$ is a discrete double divergence, but its exact boundary adjoint is what defines the set. It lies in $(\ker H{)}^{\perp }$, so components in $\ker H$ are preserved by denoising. Interior second derivatives annihilate affine image signals; whether all such image signals remain in the kernel depends on the boundary convention. With a pixelwise Hessian norm, the corresponding balls would instead be four-component blocks.

Use the scaled proximal operator ${\mathrm{prox}}_{\gamma E}$; take $\gamma =1$ for each unscaled resolvent.

For the quadratic case, write $D^{T}D=\mathrm{diag}(d_1,\dots ,d_n)$ and $\xi =Qx$. Since $Q$ is an orthogonal matrix, the quadratic proximal term becomes $\Vert \xi -Qz{\Vert }^2/2$. The subgradient optimality condition givesAll $d_j\ge 0$, so each diagonal entry is positive, including when $d_j=0$. Multiplying by the inverse diagonal matrix givesThe proximal operator of Poisson data fidelity separates pointwise. At each $t$, the scalar optimality equation on $x>0$ isSince $y>0$, the two roots have opposite signs. Only the positive root is admissible in the logarithm:The negative root is outside the effective domain. The second derivative $1+\gamma y/x^2$ of the proximal objective is positive, establishing the unique pointwise minimizer. The integral formula is understood on the function spaces where these quantities are admissible.

Finally, for $E(x)=|x|$, positive and negative minimizers obey $x-z+\gamma =0$ and $x-z-\gamma =0$, respectively. At zero the subdifferential condition is $z\in [-\gamma ,\gamma ]$. Combining these cases yields soft thresholding:Thus $\gamma =1$ supplies all three requested closed forms.

Use the standard inverse-problem setting of a bounded linear $K:{\ell }^2\to {\ell }^2$ and exact data having a finite-penalty solution. Let $u^{†}$ denote the unique $J$-minimizing exact solution for $J(u)=\Vert u{\Vert }_1+\beta \Vert u{\Vert }_2^2$, $\beta >0$. This is the limiting target of elastic net regularization in Hilbert sequence space; it need not be the Moore–Penrose minimum-norm solution unless additional hypotheses identify them, for example injectivity of $K$.

For each $\alpha >0$, the reconstruction objective is weakly lower semicontinuous, coercive by $\alpha \beta \Vert u{\Vert }_2^2$, and strictly convex. The direct method on the reflexive l2 sequence space gives existence, and strict convexity gives uniqueness. For two data sets, monotonicity of the one-norm subgradients and the quadratic penalty give, with $d=u_1-u_2$ and $g=f_1-f_2$,Completing the square yields $\Vert d\Vert \le \Vert g\Vert /(2\sqrt{2\alpha \beta })$, so each reconstruction map is continuous.

Minimality against $u^{†}$ givesChoose $\alpha \to 0$ and ${\delta }^2/\alpha \to 0$, for example $\alpha =\delta$. This bounds the Hilbert norms, drives the data residual to zero and gives $\mathrm{limsup}J(u_{\alpha }^{\delta })\le J(u^{†})$. Every weakly convergent subsequence has an exact solution as its limit, by boundedness of $K$; weak lower semicontinuity and uniqueness of the penalty minimizer identify that limit with $u^{†}$. Separately applying weak lower semicontinuity to the one-norm givesWeak convergence and convergence of Hilbert norms imply strong convergence by the Radon-Riesz theorem. This proves norm convergence of elastic net regularization:The argument applies to every sequence of noise levels and admissible data, hence gives the usual uniform noise-ball convergence for a fixed admissible exact datum.

For nonlinearity of elastic net reconstruction, zero is the minimizer exactly when $\Vert K^{\ast }{f}^{\delta }{\Vert }_{\infty }\le \alpha$. This includes a neighborhood of zero. If $K\ne 0$, choose $e_j$ with $K{e}_j\ne 0$ and $f^{\delta }=tK{e}_j$ for $t\Vert K{e}_j{\Vert }^2>\alpha$; zero then fails its optimality condition. A linear map vanishing on a neighborhood would vanish everywhere, so the elastic-net reconstruction is nonlinear whenever $K\ne 0$. For $K=I$ the explicit formula is $[R_{\alpha }f{]}_j=\mathrm{sign}(f_j)(|f_j|-\alpha {)}_{+}/(1+2\alpha \beta )$, exhibiting soft thresholding.

A concrete distinction from Moore–Penrose convergence is $K(u)=u_1+2u_2$ (embedded in the first output coordinate of ${\ell }^2$), exact datum one, and $\beta =1$. Its minimum-norm solution is $(1/5,2/5,0,\dots )$, with penalty $4/5$, while the unique penalty-minimizing exact solution is $(0,1/2,0,\dots )$, with penalty $3/4$. For $0<\alpha <2$ the exact-data elastic-net reconstruction is $(0,(2-\alpha )/(4+2\alpha ),0,\dots )$, converging to the latter. Thus it is not a Moore–Penrose-consistent regularization for every noninjective operator under the earlier definition.

The PDF supplies neither a nonzero assumption on $K$ nor a target/domain qualification here. Its unconditional nonlinearity claim is false for $K=0$, whose reconstruction is the zero linear map. The convergence proof above states the finite-penalty exact-solution hypotheses explicitly; it establishes convergence to the penalty-minimizing solution, not an automatically identical minimum-norm least-squares solution.

For a differentiable convex function $F$ with $L$-Lipschitz gradient and a proper lower-semicontinuous convex function $H$, the proximal gradient method isThe proximal operator is ${\mathrm{prox}}_{\tau H}(v)=\arg {\min }_z\{H(z)+\Vert z-v{\Vert }^2/(2\tau )\}$. To obtain an explicit iteration for discrete isotropic total variation, let $D=\nabla$ with the printed zero boundary differences andThe dual is minimization of $F(p)=\frac{1}{2}\Vert g-D^{\ast }p{\Vert }_2^2$ over $\mathcal{P}$, and the primal minimizer is $u_{\ast }=g-D^{\ast }{p}_{\ast }$. Here $D^{\ast }=-\mathrm{div}$ must be the exact adjoint operator of these differences. For example,with the analogous formula in $j$ for $(D_y^{+}{)}^{\ast }$; for $N=1$ both operators are zero. Equivalently use $p_{0j}^1=p_{Nj}^1=p_{i0}^2=p_{iN}^2=0$ in the backward-difference expression, ignoring unused edge components.

The gradient of the dual objective is $D(D^{\ast }p-g)$, with Lipschitz constant $\Vert D{\Vert }^2\le 8$. Taking $H$ to be the indicator functional of a constraint set makes its proximal operator the Euclidean projection onto a convex set. Thus a valid explicit algorithm isThe bound follows by summing $(a-b{)}^2\le 2a^2+2b^2$ over grid differences. Finite-dimensional proximal gradient method convergence gives a dual minimizer and convergence of the recovered $u^k$ to the unique primal minimizer. Projection is pointwise in the dual; applying pointwise soft thresholding directly to $Du$ would not compute the primal TV proximal operator.

Take $F(x)=\frac{1}{2}\Vert Ax-y{\Vert }_2^2$, with $\nabla F(x)=A^T(Ax-y)$ and Lipschitz constant $L=\Vert A{\Vert }_{2\to 2}^2$, and $H(x)=\Vert x{\Vert }_1$. The proximal operator separates over coordinates. Solving the scalar subgradient optimality condition $0\in \partial |z|+(z-v)/\tau$ gives soft thresholding $S_{\tau }(v)=\mathrm{sgn}(v)(|v|-\tau {)}_{+}$. Therefore the iterative soft-thresholding algorithm isHere soft thresholding is applied coordinatewise. The objective is convex and coercive because of its ${\ell }^1$ term, so minimizers exist and the proximal gradient method converges to a minimizer. Uniqueness need not hold if $A$ has a nontrivial null space. If $A=0$, the unique minimizer is zero and any positive step size is admissible.

A vector $g$ is a subgradient of a convex function $f$ at $x$ whenA point $x$ minimizes $f$ exactly when $0\in \partial f(x)$. For absolute value,Coordinate descent repeatedly minimizes the objective over one coordinate while holding the others fixed, cycling through coordinates until convergence.

The Lasso solvesFor the first update, define the partial residual and scoreSince $\Vert X_1{\Vert }_2^2=n$, the coordinate objective differs by a constant fromIts subgradient condition gives the soft thresholding update

For the stated Berhu penalty, the derivative is $\mathrm{sgn}(t)$ when $0<|t|\le \delta$ and $t/\delta$ when $|t|>\delta$. The coordinatewise Karush-Kuhn-Tucker conditions therefore give

When $X^{T}X=I_p$, expanding the objective separates it by coordinates:The soft thresholding solution is therefore

For residual $r=y_i-x_i^{\top }\theta$, minimizeover $\gamma$. Soft thresholding gives $\hat{\gamma }=\mathrm{sgn}(r)(|r|-k\sigma {)}_{+}$, and the minimized value isThis is $\sigma {\rho }_k(r/\sigma )$ for the usual Huber loss. Multiplication by the positive constant $\sigma$ does not change the minimizing $\theta$, proving equivalence.

Hard thresholding keeps detail coefficients whose magnitude exceeds a prescribed threshold and sets the rest to zero, while retaining the coarse approximation coefficients. Soft thresholding instead uses $\mathrm{sgn}(\hat{b})(|\hat{b}|-\tau {)}_{+}$. The simultaneous wavelet coefficient noise bound motivates thresholds of order $\sigma \sqrt{\log n/n}$ in a normalized Gaussian regression model with at most order $n$ details. This is a nonlinear selection rule: coefficients can be removed individually rather than discarding an entire finest resolution.