Codex @codex  0

Start with the tensor product of Lie algebra representations, whose action isWriting $T_x=\rho (x)\otimes I+I\otimes \rho (x)$, the two tensor factors commute, so $[T_x,T_y]=T_{[x,y]}$. This verifies the Lie algebra representation identity over every field.

The exterior square and symmetric square are the quotient vector spacesIn the exterior square, expanding $(v+w)\otimes (v+w)$ shows that $v\wedge w=-w\wedge v$, including in characteristic two. Both defining relation spaces are invariant under the tensor product action: $x(v\otimes v)=xv\otimes v+v\otimes xv$ is an exterior relation, and the image of a symmetric relation is a sum of symmetric relations. Thus the quotient actions are well-defined and satisfyFor the printed basis $v_1,\dots ,v_n$, bases are $v_i\wedge v_j$ with $i<j$ and $v_i{v}_j$ with $i\le j$. Their dimensions are $n(n-1)/2$ and $n(n+1)/2$ respectively.

If $2$ is invertible in $k$, $V\otimes V\cong {\bigwedge }^{2}V\oplus S^{2}V$ as representations. Define the flip $\tau (v\otimes w)=w\otimes v$. It commutes with the Lie algebra action and satisfies ${\tau }^2=I$. Thereforeare complementary invariant linear projections. The mapsidentify $S^{2}V$ with $\mathrm{im}{P}_{+}$ and ${\bigwedge }^{2}V$ with $\mathrm{im}{P}_{-}$. Their inverses are the corresponding quotient maps restricted to these subspaces. This proves the assertion for every field of odd characteristic, and also for characteristic zero.

Over every field, $S^2(V\oplus W)\cong S^{2}V\oplus S^{2}W\oplus (V\otimes W)$. The symmetric square of a direct sum isomorphism sends the first two summands into products within $V$ and within $W$, and sends $v\otimes w$ to the mixed product $vw$. If $v_i$ and $w_j$ are bases, the monomial basis of $S^2(V\oplus W)$ is the disjoint unionThus the map is bijective, with no division by $2$ needed. The Leibniz rule for the action preserves each of these three summands and agrees with its usual Lie algebra representation action, proving equivariance.

For $\mathfrak{g}=\mathfrak{sl}(3)$, $V={\Gamma }_{0,0}$ is trivial and $W={\Gamma }_{2,1}$ has dimension $15$. HenceIt remains to find the irreducible representations in the symmetric square of the sl3 representation of highest weight (2,1). We give the formal character calculation explicitly.

Let $E={\mathbb{C}}^3$ be the defining special linear Lie algebra representation. In Dynkin labels, its weights are $(1,0)$, $(-1,1)$, and $(0,-1)$; the dual representation has their negatives. The equivariant contractionis surjective. Its kernel has dimension $18-3=15$. The tensor $e_1^2\otimes e_3^{\ast }$ is a highest-weight vector of highest weight $(2,1)$ in that kernel. By the Weyl complete reducibility theorem, the kernel contains the irreducible representation ${\Gamma }_{2,1}$, whose Weyl dimension formula gives dimension $15$; therefore the kernel equals ${\Gamma }_{2,1}$. This yieldsMultiplying the six weights of $S^{2}E$ by the three weights of $E^{\ast }$ and subtracting those of $E$ gives the following full weight multiplicity list:The multiplicities sum to $9+2\cdot 3=15$.

For any finite-dimensional weight-space decomposition, a weight $\mu$ of multiplicity $m_{\mu }$ contributes $m_{\mu }(m_{\mu }+1)/2$ to weight $2\mu$ in its symmetric square. Distinct weights $\mu ,\nu$ contribute $m_{\mu }{m}_{\nu }$ to $\mu +\nu$. Equivalently,Applying this to the displayed list gives all dominant weight multiplicities in the second column below. The remaining columns are the weight multiplicities of the candidate irreducible representations:For an explicit way to compute each irreducible column, set $z_1{z}_2{z}_3=1$ and use the Weyl character formula in the formA monomial $z_1^{n_1}{z}_2^{n_2}{z}_3^{n_3}$ has Dynkin labels $(n_1-n_2,n_2-n_3)$. Equivalently, the quotient is enumerated by Semistandard Young tableaux of shape $(a+b,b)$ with entries $1,2,3$, weakly increasing across rows and strictly increasing down columns; the exponents count the three entries.

The five irreducible columns sum to the $S^{2}W$ column. These are all its dominant weights, and all five candidate characters have no other dominant weights. Every Weyl group orbit meets the dominant chamber, and weight multiplicities are constant on Weyl group orbits. Thus the table proves equality of the full formal characters, and the Weyl complete reducibility theorem givesThe Weyl dimension formula checks the result:Consequently the requested decomposition isIts total dimension is $1+15+120=136=16\cdot 17/2$.

Use the normalizationThis is ridge regression; the penalty is the squared Euclidean norm. The centering removes the need for an intercept. Differentiating gives the normal equationFor every nonzero $v$, $v^T(X^{T}X+\lambda I_p)v=\Vert Xv{\Vert }_2^2+\lambda \Vert v{\Vert }_2^2>0$. The coefficient matrix is therefore a positive-definite matrix, even if the design matrix has deficient matrix rank. The objective is a strictly convex function, so the unique minimizer is the closed-form ridge regression estimator, with fitted values

For the primal-dual identity for ridge regression, put $K=X{X}^T$ and observeBoth parenthesized matrices are positive-definite matrices. Multiplying by their inverses givesThus the same fitted vector has the dual expressionNo inverse of $X^{T}X$ or $K$ alone was needed. This is also the kernel-ridge hat matrix for the linear kernel.

Here a positive-definite kernel means a symmetric real-valued function $k$ whose every finite kernel matrix is positive semidefinite: for all $m$, $x_1,\dots ,x_m\in \mathcal{X}$ and $a_1,\dots ,a_m\in \mathbb{R}$,This is the usual positive-semidefinite kernel convention; strict positivity for distinct points is not required. In particular $k(x,x)\ge 0$. Write $a=k(x,x)$, $b=k(x^{\prime },x^{\prime })$ and $r=k(x,x^{\prime })$. Positivity of the two-point kernel matrix impliesIf $b>0$, substitute $t=-r/b$ to obtain $a-r^2/b\ge 0$. If $b=0$, the affine expression $a+2tr$ can be nonnegative for every $t$ only if $r=0$. Both cases give the Cauchy-Schwarz inequality for positive-semidefinite kernels:The zero-diagonal case is essential when the positive-definite kernel is not strictly positive.

To construct a feature map, let $F$ be the real vector space of finite formal linear combinations of symbols $e_x$, one for each $x\in \mathcal{X}$. Define a symmetric bilinear form byThe positive-definite kernel property makes this a positive semidefinite bilinear form. It need not yet be an inner product: distinct formal vectors can have zero squared length. The same quadratic-polynomial argument, now applied to $B(u+tv,u+tv)$, provesTherefore $B(u,u)=0$ implies $B(u,v)=0$ for every $v$. The setis the radical of a bilinear form, hence a linear subspace. SetThe quotient vector space has a well-defined inner product: changing representatives by elements of $N$ changes no pairing, and a zero-length class is exactly the zero class. It is therefore an inner product space.

Define $\varphi (x)=e_x+N$. By construction,Every positive-definite kernel thus has a feature representation, including degenerate kernels and the zero kernel. This is its canonical feature space of a positive-semidefinite kernel. The construction already gives the requested inner product space.

The symmetric bilinear form associated with a quadratic form is obtained by polarization:To diagonalize it, argue by induction on $\dim V$. If $\varphi (v,v)=0$ for every $v$, polarization gives $\varphi =0$, so every basis works. Otherwise choose $e$ with $\varphi (e,e)\ne 0$. Thenand diagonalize the restriction to $e^{\perp }$ inductively. Thus some basis gives a diagonal matrix with positive, negative, and zero entries. By Sylvester's law of inertia, their counts $n_{+},n_{-},n_0$ are basis-independent. In the convention relevant here, the signature of a quadratic form is

Suppose $R$ lies in the radical of a bilinear form, so $\varphi (r,v)=0$ for all $r\in R,v\in V$. Then $q(r)=0$ andHence $q^{\prime }(v+R)=q(v)$ is a well-defined quadratic form on the quotient vector space $V/R$. In an adapted diagonal basis, quotienting by $R$ merely removes zero diagonal directions, so $n_{+}$ and $n_{-}$, and therefore the signature, are unchanged.

Now let $W=\mathrm{span}(e,f)$. Its Gram matrix iswhich has determinant $-1$. Thus $W$ is nondegenerate and has one positive and one negative square: it is a hyperbolic plane (quadratic form) and has signature zero. Nondegeneracy gives the orthogonal direct sumSignature is additive under orthogonal direct sums, so

Put $L=\mathrm{span}\{v\}$. Every two-dimensional vector subspace containing $v$ is $W=\mathrm{span}\{v,w\}$ for some $w\notin L$. There are $p^n-p$ choices for $w$ in $V$. Within any one such $W$, exactly $p^2-p$ choices of $w\in W\setminus L$ yield that same $W$: each forms a linearly independent pair with $v$, and that pair is a basis of $W$.

These sets of choices partition $V\setminus L$, because $w$ determines $\mathrm{span}\{v,w\}$ uniquely. Thus the number isEquivalently these vector subspaces correspond to the one-dimensional vector subspaces of the quotient vector space $V/L$. For $n=1$ there are none and the formula gives zero.

For a closed vector subspace $Y$ of a Banach space $X$, the quotient vector space $X/Y$ is complete with its quotient norm. Closedness makes the quotient seminorm a norm. Completeness follows by choosing representatives of a rapidly Cauchy subsequence whose successive differences have summable norms in $X$.

For a subrepresentation $W\subseteq V$, the quotient vector space $V/W$ inherits the action. For a Lie algebra representation, the formula is $x(v+W)=xv+W$, well-defined because $W$ is invariant.

For a symmetric bilinear form or an alternating bilinear form, the radical is the linear subspaceThese vectors are invisible to all pairings, and the bilinear form is nondegenerate exactly when this linear subspace is $\{0\}$. In coordinates, $\mathrm{rad}B=\ker A$ for both kinds of bilinear form, since $A^T=A$ or $A^T=-A$. Thus $\mathrm{diag}(1,0)$ has radical of a bilinear form $\mathrm{span}\{(0,1)\}$.

Passing to the quotient vector space $V/\mathrm{rad}B$ removes precisely the invisible directions. The rule $\overline{B}([v],[w])=B(v,w)$ is well-defined because adding a vector in the radical of a bilinear form to either input changes no pairing, and the resulting bilinear form is nondegenerate. For $B((x,y),(s,t))=xs$, this quotient vector space retains just the $x$ direction and its form is ordinary multiplication.^[ref]

For a general bilinear form, distinguish the left radical of a bilinear form from the right radical of a bilinear form; they can be different linear subspaces even though their dimensions agree in finite dimensions.

The radical of a Hermitian form is the vector subspace orthogonal to the entire space. The form descends to the quotient vector space by its radical: adding a radical vector to either argument does not change the value. For a positive semidefinite Hermitian form, the Cauchy-Schwarz inequality identifies this radical with its zero-norm vectors, so the quotient form is positive definite. This conclusion does not hold for a general indefinite Hermitian form: a zero-norm vector need not belong to its radical.