This paper concerns the use of a particular class of determinantal point processes (DPP), a class of repulsive spatial point processes, for Monte Carlo integration. Let $d\ge 1$, $I\subseteq \overline d=\{1,\dots,d\}$ with $\iota=|I|$. Using a single set of $N$ quadrature points $\{u_1,\dots,u_N\}$ defined, once for all, in dimension $d$ from the realization of the DPP model, we investigate "minimal" assumptions on the integrand in order to obtain unbiased Monte Carlo estimates of $\mu(f_I)=\int_{[0,1]^\iota} f_I(u) \mathrm{d} u$ for any known $\iota$-dimensional integrable function on $[0,1]^\iota$. In particular, we show that the resulting estimator has variance with order $N^{-1-(2s\wedge 1)/d}$ when the integrand belongs to some Sobolev space with regularity $s > 0$. When $s>1/2$ (which includes a large class of non-differentiable functions), the variance is asymptotically explicit and the estimator is shown to satisfy a Central Limit Theorem.
相關內容
We explore a new way to handle flux boundary conditions imposed on level sets. The proposed approach is a diffuse interface version of the shifted boundary method (SBM) for continuous Galerkin discretizations of conservation laws in embedded domains. We impose the interface conditions weakly and approximate surface integrals by volume integrals. The discretized weak form of the governing equation has the structure of an immersed boundary finite element method. A ghost penalty term is included to extend the weak solution into the external subdomain. The calculation of interface forcing terms requires (i) construction of an approximate delta function and (ii) extrapolation of embedded boundary data into quadrature points. We accomplish these tasks using a level set function, which is given analytically or evolved numerically. A globally defined averaged gradient of this approximate signed distance function is used to construct a simple map to the closest point on the interface. The normal and tangential derivatives of the numerical solution at that point are calculated using the interface conditions and/or interpolation on uniform stencils. Similarly to SBM, extrapolation back to the quadrature points is performed using Taylor expansions. The same strategy is used to construct ghost penalty functions and extension velocities. Computations that require extrapolation are restricted to a narrow band around the interface. Numerical results are presented for elliptic, parabolic, and hyperbolic test problems, which are specifically designed to assess the error caused by the numerical treatment of interface conditions on fixed and moving boundaries in 2D.
We study the mathematical structure of the solution set (and its tangent space) to the matrix equation $G^*JG=J$ for a given square matrix $J$. In the language of pure mathematics, this is a Lie group which is the isometry group for a bilinear (or a sesquilinear) form. Generally these groups are described as intersections of a few special groups. The tangent space to $\{G: G^*JG=J \}$ consists of solutions to the linear matrix equation $X^*J+JX=0$. For the complex case, the solution set of this linear equation was computed by De Ter{\'a}n and Dopico. We found that on its own, the equation $X^*J+JX=0$ is hard to solve. By throwing into the mix the complementary linear equation $X^*J-JX=0$, we find that rather than increasing the complexity, we reduce the complexity. Not only is it possible to now solve the original problem, but we can approach the broader algebraic and geometric structure. One implication is that the two equations form an $\mathfrak{h}$ and $\mathfrak{m}$ pair familiar in the study of pseudo-Riemannian symmetric spaces. We explicitly demonstrate the computation of the solutions to the equation $X^*J\pm XJ=0$ for real and complex matrices. However, any real, complex or quaternionic case with an arbitrary involution (e.g., transpose, conjugate transpose, and the various quaternion transposes) can be effectively solved with the same strategy. We provide numerical examples and visualizations.
An efficient implicit representation of an $n$-vertex graph $G$ in a family $\mathcal{F}$ of graphs assigns to each vertex of $G$ a binary code of length $O(\log n)$ so that the adjacency between every pair of vertices can be determined only as a function of their codes. This function can depend on the family but not on the individual graph. Every family of graphs admitting such a representation contains at most $2^{O(n\log(n))}$ graphs on $n$ vertices, and thus has at most factorial speed of growth. The Implicit Graph Conjecture states that, conversely, every hereditary graph family with at most factorial speed of growth admits an efficient implicit representation. We refute this conjecture by establishing the existence of hereditary graph families with factorial speed of growth that require codes of length $n^{\Omega(1)}$.
Discretization of the uniform norm of functions from a given finite dimensional subspace of continuous functions is studied. We pay special attention to the case of trigonometric polynomials with frequencies from an arbitrary finite set with fixed cardinality. We give two different proofs of the fact that for any $N$-dimensional subspace of the space of continuous functions it is sufficient to use $e^{CN}$ sample points for an accurate upper bound for the uniform norm. Previous known results show that one cannot improve on the exponential growth of the number of sampling points for a good discretization theorem in the uniform norm. Also, we prove a general result, which connects the upper bound on the number of sampling points in the discretization theorem for the uniform norm with the best $m$-term bilinear approximation of the Dirichlet kernel associated with the given subspace. We illustrate application of our technique on the example of trigonometric polynomials.
A Boolean maximum constraint satisfaction problem, Max-CSP($f$), is specified by a constraint function $f:\{-1,1\}^k\to\{0,1\}$; an instance on $n$ variables is given by a list of constraints applying $f$ on a tuple of "literals" of $k$ distinct variables chosen from the $n$ variables. Chou, Golovnev, and Velusamy [CGV20] obtained explicit constants characterizing the streaming approximability of all symmetric Max-2CSPs. More recently, Chou, Golovnev, Sudan, and Velusamy [CGSV21] proved a general dichotomy theorem tightly characterizing the approximability of Boolean Max-CSPs with respect to sketching algorithms. For every $f$, they showed that there exists an optimal approximation ratio $\alpha(f)\in (0,1]$ such that for every $\epsilon>0$, Max-CSP($f$) is $(\alpha(f)-\epsilon)$-approximable by a linear sketching algorithm in $O(\log n)$ space, but any $(\alpha(f)+\epsilon)$-approximation sketching algorithm for Max-CSP($f$) requires $\Omega(\sqrt{n})$ space. In this work, we build on the [CGSV21] dichotomy theorem and give closed-form expressions for the sketching approximation ratios of multiple families of symmetric Boolean functions. The functions include $k$AND and Th$_k^{k-1}$ (the ``weight-at-least-$(k-1)$'' threshold function on $k$ variables). In particular, letting $\alpha'_k = 2^{-(k-1)} (1-k^{-2})^{(k-1)/2}$, we show that for odd $k \geq 3$, $\alpha(k$AND$ = \alpha'_k$; for even $k \geq 2$, $\alpha(k$AND$) = 2\alpha'_{k+1}$; and for even $k \geq 2$, $\alpha($Th$_k^{k-1}) = \frac{k}2\alpha'_{k-1}$. We also resolve the ratio for the ``weight-exactly-$\frac{k+1}2$'' function for odd $k \in \{3,\ldots,51\}$ as well as fifteen other functions. These closed-form expressions need not have existed just given the [CGSV21] dichotomy. For arbitrary threshold functions, we also give optimal "bias-based" approximation algorithms generalizing [CGV20] and simplifying [CGSV21].
Let $f$ be analytic on $[0,1]$ with $|f^{(k)}(1/2)|\leq A\alpha^kk!$ for some constant $A$ and $\alpha<2$. We show that the median estimate of $\mu=\int_0^1f(x)\,\mathrm{d}x$ under random linear scrambling with $n=2^m$ points converges at the rate $O(n^{-c\log(n)})$ for any $c< 3\log(2)/\pi^2\approx 0.21$. We also get a super-polynomial convergence rate for the sample median of $2k-1$ random linearly scrambled estimates, when $k=\Omega(m)$. When $f$ has a $p$'th derivative that satisfies a $\lambda$-H\"older condition then the median-of-means has error $O( n^{-(p+\lambda)+\epsilon})$ for any $\epsilon>0$, if $k\to\infty$ as $m\to\infty$.
In 2006, Arnold, Falk, and Winther developed finite element exterior calculus, using the language of differential forms to generalize the Lagrange, Raviart-Thomas, Brezzi-Douglas-Marini, and N\'ed\'elec finite element spaces for simplicial triangulations. In a recent paper, Licht asks whether, on a single simplex, one can construct bases for these spaces that are invariant with respect to permuting the vertices of the simplex. For scalar fields, standard bases all have this symmetry property, but for vector fields, this question is more complicated: such invariant bases may or may not exist, depending on the polynomial degree of the element. In dimensions two and three, Licht constructs such invariant bases for certain values of the polynomial degree $r$, and he conjectures that his list is complete, that is, that no such basis exists for other values of $r$. In this paper, we show that Licht's conjecture is true in dimension two. However, in dimension three, we show that Licht's ideas can be extended to give invariant bases for many more values of $r$; we then show that this new larger list is complete. Along the way, we develop a more general framework for the geometric decomposition ideas of Arnold, Falk, and Winther.
A T-graph (a special case of a chordal graph) is the intersection graph of connected subtrees of a suitable subdivision of a fixed tree T . We deal with the isomorphism problem for T-graphs which is GI-complete in general - when T is a part of the input and even a star. We prove that the T-graph isomorphism problem is in FPT when T is the fixed parameter of the problem. This can equivalently be stated that isomorphism is in FPT for chordal graphs of (so-called) bounded leafage. While the recognition problem for T-graphs is not known to be in FPT wrt. T, we do not need a T-representation to be given (a promise is enough). To obtain the result, we combine a suitable isomorphism-invariant decomposition of T-graphs with the classical tower-of-groups algorithm of Babai, and reuse some of the ideas of our isomorphism algorithm for S_d-graphs [MFCS 2020].
Inversion of the two-dimensional discrete Fourier transform (DFT) typically requires all DFT coefficients to be known. When only band-limited DFT coefficients of a matrix are known, the original matrix can not be uniquely recovered. Using a priori information that the matrix is binary (all elements are either 0 or 1) can overcome the missing high-frequency DFT coefficients and restore uniqueness. We theoretically investigate the smallest pass band that can be applied while still guaranteeing unique recovery of an arbitrary binary matrix. The results depend on the dimensions of the matrix. Uniqueness results are proven for the dimensions $p\times q$, $p\times p$, and $p^\alpha\times p^\alpha$, where $p\neq q$ are primes numbers and $\alpha>1$ an integer. An inversion algorithm is proposed for practically recovering the unique binary matrix. This algorithm is based on integer linear programming methods and significantly outperforms naive implementations. The algorithm efficiently reconstructs $17\times17$ binary matrices using 81 out of the total 289 DFT coefficients.
We show that if $f\colon S_n \to \{0,1\}$ is $\epsilon$-close to linear in $L_2$ and $\mathbb{E}[f] \leq 1/2$ then $f$ is $O(\epsilon)$-close to a union of "mostly disjoint" cosets, and moreover this is sharp: any such union is close to linear. This constitutes a sharp Friedgut-Kalai-Naor theorem for the symmetric group. Using similar techniques, we show that if $f\colon S_n \to \mathbb{R}$ is linear, $\Pr[f \notin \{0,1\}] \leq \epsilon$, and $\Pr[f = 1] \leq 1/2$, then $f$ is $O(\epsilon)$-close to a union of mostly disjoint cosets, and this is also sharp; and that if $f\colon S_n \to \mathbb{R}$ is linear and $\epsilon$-close to $\{0,1\}$ in $L_\infty$ then $f$ is $O(\epsilon)$-close in $L_\infty$ to a union of disjoint cosets.
小貼士
登錄享
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191