Search

The Polyak-Lojasiewicz (PL) constant of a function $f \colon \mathbb{R}^d \to \mathbb{R}$ characterizes the best exponential rate of convergence of gradient flow for $f$, uniformly over initializations. Meanwhile, in the theory of Markov diffusions, the log-Sobolev (LS) constant plays an analogous role, governing the exponential rate of convergence for the Langevin dynamics from arbitrary initialization in the Kullback-Leibler divergence. We establish a new connection between optimization and sampling by showing that the low temperature limit $\lim_{t\to 0^+} t^{-1} C_{\mathsf{LS}}(\mu_t)$ of the LS constant of $\mu_t \propto \exp(-f/t)$ is exactly the PL constant of $f$, under mild assumptions. In contrast, we show that the corresponding limit for the Poincar\'e constant is the inverse of the smallest eigenvalue of $\nabla^2 f$ at the minimizer., Comment: 22 pages

Geometric tempering is a popular approach to sampling from challenging multi-modal probability distributions by instead sampling from a sequence of distributions which interpolate, using the geometric mean, between an easier proposal distribution and the target distribution. In this paper, we theoretically investigate the soundness of this approach when the sampling algorithm is Langevin dynamics, proving both upper and lower bounds. Our upper bounds are the first analysis in the literature under functional inequalities. They assert the convergence of tempered Langevin in continuous and discrete-time, and their minimization leads to closed-form optimal tempering schedules for some pairs of proposal and target distributions. Our lower bounds demonstrate a simple case where the geometric tempering takes exponential time, and further reveal that the geometric tempering can suffer from poor functional inequalities and slow convergence, even when the target distribution is well-conditioned. Overall, our results indicate that geometric tempering may not help, and can even be harmful for convergence.

Motivated by the manifold hypothesis, which states that data with a high extrinsic dimension may yet have a low intrinsic dimension, we develop refined statistical bounds for entropic optimal transport that are sensitive to the intrinsic dimension of the data. Our bounds involve a robust notion of intrinsic dimension, measured at only a single distance scale depending on the regularization parameter, and show that it is only the minimum of these single-scale intrinsic dimensions which governs the rate of convergence. We call this the Minimum Intrinsic Dimension scaling (MID scaling) phenomenon, and establish MID scaling with no assumptions on the data distributions so long as the cost is bounded and Lipschitz, and for various entropic optimal transport quantities beyond just values, with stronger analogs when one distribution is supported on a manifold. Our results significantly advance the theoretical state of the art by showing that MID scaling is a generic phenomenon, and provide the first rigorous interpretation of the statistical effect of entropic regularization as a distance scale., Comment: 53 pages

We study the sample complexity of entropic optimal transport in high dimensions using computationally efficient plug-in estimators. We significantly advance the state of the art by establishing dimension-free, parametric rates for estimating various quantities of interest, including the entropic regression function which is a natural analog to the optimal transport map. As an application, we propose a practical model for transfer learning based on entropic optimal transport and establish parametric rates of convergence for nonparametric regression and classification., Comment: 28 pages

Sunscreen lotions are used to protect the skin from damage due to solar ultraviolet (UV) radiation. The active UV blocking components can be organic molecules or inorganic particles, for example TiO2. While both in vivo and in vitro methods exist for assessing the protective capacity of sunscreens, their basic optical properties have received little attention. In this paper we take the first steps to address this issue by presenting a method to obtain spectra of absorption and scattering coefficients of a sunscreen by analysis of experimental transmittance and reflectance spectra. We model the optical properties of the sunscreen by two flux theory, develop approximate expressions for interface reflection coefficients and obtain the remaining optical parameters by fitting the two-flux theory to measurements. We analyze spectrophotometric reflectance and transmittance spectra of experimental narrow band UV blocking materials based on mixtures of mesoporous magnesium carbonate and TiO2, as well as some commercial sunscreens. Our method is demonstrated in vitro using quartz substrates and is further used to model in vivo measurements on the upside of the hand of one of the authors. The obtained absorption and scattering coefficient spectra give information on the relative contributions of these processes to the UV blocking effect as well as the scattering of visible light responsible for the whitening of the skin upon application of a sunscreen lotion. Our analysis method may be useful for computational optimization of novel sunscreen formulations as well as for the development of in vivo test methods., Comment: 21 pages, 8 figures

We compute exact second-order asymptotics for the cost of an optimal solution to the entropic optimal transport problem in the continuous-to-discrete, or semi-discrete, setting. In contrast to the discrete-discrete or continuous-continuous case, we show that the first-order term in this expansion vanishes but the second-order term does not, so that in the semi-discrete setting the difference in cost between the unregularized and regularized solution is quadratic in the inverse regularization parameter, with a leading constant that depends explicitly on the value of the density at the points of discontinuity of the optimal unregularized map between the measures. We develop these results by proving new pointwise convergence rates of the solutions to the dual problem, which may be of independent interest.

We study first-order optimization algorithms for computing the barycenter of Gaussian distributions with respect to the optimal transport metric. Although the objective is geodesically non-convex, Riemannian GD empirically converges rapidly, in fact faster than off-the-shelf methods such as Euclidean GD and SDP solvers. This stands in stark contrast to the best-known theoretical results for Riemannian GD, which depend exponentially on the dimension. In this work, we prove new geodesic convexity results which provide stronger control of the iterates, yielding a dimension-free convergence rate. Our techniques also enable the analysis of two related notions of averaging, the entropically-regularized barycenter and the geometric median, providing the first convergence guarantees for Riemannian GD for these problems., Comment: v3: Fixed error in Theorem 3, see footnote 2

We propose a new method for smoothly interpolating probability measures using the geometry of optimal transport. To that end, we reduce this problem to the classical Euclidean setting, allowing us to directly leverage the extensive toolbox of spline interpolation. Unlike previous approaches to measure-valued splines, our interpolated curves (i) have a clear interpretation as governing particle flows, which is natural for applications, and (ii) come with the first approximation guarantees on Wasserstein space. Finally, we demonstrate the broad applicability of our interpolation methodology by fitting surfaces of measures using thin-plate splines., Comment: 38 pages, 5 figures

Motivated by the problem of sampling from ill-conditioned log-concave distributions, we give a clean non-asymptotic convergence analysis of mirror-Langevin diffusions as introduced in Zhang et al. (2020). As a special case of this framework, we propose a class of diffusions called Newton-Langevin diffusions and prove that they converge to stationarity exponentially fast with a rate which not only is dimension-free, but also has no dependence on the target distribution. We give an application of this result to the problem of sampling from the uniform distribution on a convex body using a strategy inspired by interior-point methods. Our general approach follows the recent trend of linking sampling and optimization and highlights the role of the chi-squared divergence. In particular, it yields new results on the convergence of the vanilla Langevin diffusion in Wasserstein distance., Comment: 27 pages, 10 figures

We study first order methods to compute the barycenter of a probability distribution $P$ over the space of probability measures with finite second moment. We develop a framework to derive global rates of convergence for both gradient descent and stochastic gradient descent despite the fact that the barycenter functional is not geodesically convex. Our analysis overcomes this technical hurdle by employing a Polyak-Lojasiewicz (PL) inequality and relies on tools from optimal transport and metric geometry. In turn, we establish a PL inequality when $P$ is supported on the Bures-Wasserstein manifold of Gaussian probability measures. It leads to the first global rates of convergence for first order methods in this context., Comment: 28 pages, 5 figures, to appear in COLT 2020

With the introduction of the graph-theoretic time-inconsistent planning model due to Kleinberg and Oren, it has been possible to investigate the computational complexity of how a task designer best can support a present-biased agent in completing the task. In this paper, we study the complexity of finding a choice reduction for the agent; that is, how to remove edges and vertices from the task graph such that a present-biased agent will remain motivated to reach his target even for a limited reward. While this problem is NP-complete in general, this is not necessarily true for instances which occur in practice, or for solutions which are of interest to task designers. For instance, a task designer may desire to find the best task graph which is not too complicated. We therefore investigate the problem of finding simple motivating subgraphs. These are structures where the agent will modify his plan at most $k$ times along the way. We quantify this simplicity in the time-inconsistency model as a structural parameter: The number of branching vertices (vertices with out-degree at least $2$) in a minimal motivating subgraph. Our results are as follows: We give a linear algorithm for finding an optimal motivating path, i.e. when $k=0$. On the negative side, we show that finding a simple motivating subgraph is NP-complete even if we allow only a single branching vertex --- revealing that simple motivating subgraphs are indeed hard to find. However, we give a pseudo-polynomial algorithm for the case when $k$ is fixed and edge weights are rationals, which might be a reasonable assumption in practice., Comment: An extended abstract of this paper is accepted at AAAI 2020

This work establishes fast rates of convergence for empirical barycenters over a large class of geodesic spaces with curvature bounds in the sense of Alexandrov. More specifically, we show that parametric rates of convergence are achievable under natural conditions that characterize the bi-extendibility of geodesics emanating from a barycenter. These results largely advance the state-of-the-art on the subject both in terms of rates of convergence and the variety of spaces covered. In particular, our results apply to infinite-dimensional spaces such as the 2-Wasserstein space, where bi-extendibility of geodesics translates into regularity of Kantorovich potentials.

A partial complement of the graph $G$ is a graph obtained from $G$ by complementing all the edges in one of its induced subgraphs. We study the following algorithmic question: for a given graph $G$ and graph class $\mathcal{G}$, is there a partial complement of $G$ which is in $\mathcal{G}$? We show that this problem can be solved in polynomial time for various choices of the graphs class $\mathcal{G}$, such as bipartite, degenerate, or cographs. We complement these results by proving that the problem is NP-complete when $\mathcal{G}$ is the class of $r$-regular graphs.

We generalize the family of $(\sigma, \rho)$-problems and locally checkable vertex partition problems to their distance versions, which naturally captures well-known problems such as distance-$r$ dominating set and distance-$r$ independent set. We show that these distance problems are XP parameterized by the structural parameter mim-width, and hence polynomial on graph classes where mim-width is bounded and quickly computable, such as $k$-trapezoid graphs, Dilworth $k$-graphs, (circular) permutation graphs, interval graphs and their complements, convex graphs and their complements, $k$-polygon graphs, circular arc graphs, complements of $d$-degenerate graphs, and $H$-graphs if given an $H$-representation. To supplement these findings, we show that many classes of (distance) $(\sigma, \rho)$-problems are W[1]-hard parameterized by mim-width + solution size., Comment: Accepted at IPEC 2018

We investigate new graph classes of bounded mim-width, strictly extending interval graphs and permutation graphs. The graphs $K_t \boxminus K_t$ and $K_t \boxminus S_t$ are graphs obtained from the disjoint union of two cliques of size $t$, and one clique of size $t$ and one independent set of size $t$ respectively, by adding a perfect matching. We prove that : (1) interval graphs are $(K_3\boxminus S_3)$-free chordal graphs; and $(K_t\boxminus S_t)$-free chordal graphs have mim-width at most $t-1$, (2) permutation graphs are $(K_3\boxminus K_3)$-free co-comparability graphs; and $(K_t\boxminus K_t)$-free co-comparability graphs have mim-width at most $t-1$, (3) chordal graphs and co-comparability graphs have unbounded mim-width in general. We obtain several algorithmic consequences; for instance, while Minimum Dominating Set is NP-complete on chordal graphs, it can be solved in time $n^{\mathcal{O}(t)}$ on $(K_t\boxminus S_t)$-free chordal graphs. The third statement strengthens a result of Belmonte and Vatshelle stating that either those classes do not have constant mim-width or a decomposition with constant mim-width cannot be computed in polynomial time unless $P=NP$. We generalize these ideas to bigger graph classes. We introduce a new width parameter sim-width, of stronger modelling power than mim-width, by making a small change in the definition of mim-width. We prove that chordal graphs and co-comparability graphs have sim-width at most 1. We investigate a way to bound mim-width for graphs of bounded sim-width by excluding $K_t\boxminus K_t$ and $K_t\boxminus S_t$ as induced minors or induced subgraphs, and give algorithmic consequences. Lastly, we show that circle graphs have unbounded sim-width, and thus also unbounded mim-width., Comment: 24 pages, 5 figures; An extended abstract appeared in the proceedings of WALCOM2017

We propose an algebraic framework for generalized graph Laplacians which unifies the study of resistor networks, the critical group, and the eigenvalues of the Laplacian and adjacency matrices. Given a graph with boundary $G$ together with a generalized Laplacian $L$ with entries in a commutative ring $R$, we define a generalized critical group $\Upsilon_R(G,L)$. We relate $\Upsilon_R(G,L)$ to spaces of harmonic functions on the network using the Hom, Tor, and Ext functors of homological algebra. We study how these algebraic objects transform under combinatorial operations on the network $(G,L)$, including harmonic morphisms, layer-stripping, duality, and symmetry. In particular, we use layer-stripping operations from the theory of resistor networks to systematize discrete harmonic continuation. This leads to an algebraic characterization of the graphs with boundary that can be completely layer-stripped, an algorithm for simplifying computation of $\Upsilon_R(G,L)$, and upper bounds for the number of invariant factors in the critical group and the multiplicity of Laplacian eigenvalues in terms of geometric quantities., Comment: 73 pages, 27 figures. This revised version submitted to SIDMA on Jan. 9 2017

A pseudoforest is a graph whose connected components have at most one cycle. Let X be a pseudoforest modulator of graph G, i. e. a vertex subset of G such that G-X is a pseudoforest. We show that Vertex Cover admits a polynomial kernel being parameterized by the size of the pseudoforest modulator. In other words, we provide a polynomial time algorithm that for an input graph G and integer k, outputs a graph G' and integer k', such that G' has O(|X|12) vertices and G has a vertex cover of size k if and only if G' has vertex cover of size k'. We complement our findings by proving that there is no polynomial kernel for Vertex Cover parameterized by the size of a modulator to a mock forest (a graph where no cycles share a vertex) unless NP is a subset of coNP/poly. In particular, this also rules out polynomial kernels when parameterized by the size of a modulator to outerplanar and cactus graphs.

In 2014 an all-sky imager (ASI) and an Advanced Modular Incoherent Scatter Radar consisting of 14 panels (AMISR-14) system were installed at the Jicamarca Radio Observatory. The ASI measures airglow depletions associated with large-scale equatorial spread F irregularities (10s-100s km), while AMISR-14 detects small-scale irregularities (0.34 m). This study presents simultaneous observations of equatorial spread F (ESF) irregularities at 10-100 km scales using the all sky-imager, at 3 m scales using the JULIA (Jicamarca Unattended Long-term Investigations of the Ionosphere and Atmosphere) radar, and at 0.34 m scales using the AMISR-14 radar. We compare data from the three instruments on the night of 20-21 August, 2014 by locating the radar scattering volume in the optical images. During this night no topside plumes were observed, and we only compare with bottomside ESF. AMISR-14 had five beams perpendicular to the magnetic field covering ~200 km in the east-west direction at 250 km altitude. Comparing the radar data with zenith ASI measurements, we found that most of the echoes occur on the western wall of the depletions with fewer echoes observed the eastern wall and center, contrary to previous comparisons of topside plumes that showed most of the echoes in the center of depleted regions. We attribute these differences to the occurrence of irregularities produced at sub-meter scales by the lower-hybrid-drift instability. Comparisons of the ASI observations with JULIA images show similar results to those found in the AMISR-14 and ASI comparison., Comment: 35 pages (unformatted manuscript), 8 figures, 2 tables, accepted for publication in JGR Space Physics

The frog model is a system of random walks where active particles set sleeping particles in motion. On the complete graph with n vertices it is equivalent to a well-understood rumor spreading model. We given an alternate and elementary proof that the wake-up time, i.e. the expected time for every particle to be activated, is Theta(log n). Additionally, we give an explicit distributional equation for the wakeup time as a weighted sum of geometric random variables. This project was part of the University of Washington Research Experience for Undergraduates program., Comment: 9 pages

The "n! conjecture" of Garsia and Haiman has inspired mathematicians for nearly two decades, even after Haiman published a proof in 2001. Kumar and Funch Thomsen proved in 2003 that in order to prove the conjecture for all partitions, it suffices to prove it for the so-called "staircase partitions" $(k,k-1,...,2,1)$ for each $k>1$. In the present paper we give a construction of a specially designed two-dimensional family of length-$n$ subschemes of the plane, and use that to prove the $n!$ conjecture for staircase partitions. Together with the result of Kumar and Funch Thomsen, this provides a new proof of Haiman's theorem., Comment: A referee found a serious flaw in the argument, so we withdraw the paper

We study the equity-efficiency implication of giving school districts control over teacher pay using an equilibrium model of the market for public-school teachers. Teachers differ in their comparative advantages in teaching low- or high-achieving students. School districts, which serve different student bodies, use both wage and hiring strategies to compete for their preferred teachers. We estimate the model using data from Wisconsin, where districts gained control over teacher pay in 2011. We find that, all else equal, giving districts control over teacher pay would lead to more efficient teacher-district sorting but larger educational inequality. Teacher bonus programs that incentivize comparative advantage-based sorting, combined with bonus rates favoring districts with more low-achieving students, could improve both efficiency and equity.

In this article we want to present the EISCAT Svalbard Radar (ESR) in some detail, as well as some of the instruments of interest for ionospheric and magnetospheric research that are located in the vicinity of it. We particularly describe how this instrument cluster, close to the geomagnetic conjugate point of the Chinese Antarctic Zhongshan station, can contribute to inter-hemispheric coordinated studies of the polar Ionosphere., Comment: 12 pages, 3 PostScript figures

Let S be a smooth projective surface, and consider the following two subvarieties of the Hilbert scheme parameterizing closed subschemes of S of length n: A = {subschemes with support in a fixed point of S} B = {subschemes with support in one (variable) point of S} A and B have complementary dimensions in the Hilbert scheme. We prove that the intersection number [A].[B] = n(-1)^(n-1), answering a question by H. Nakajima., Comment: AMSLaTeX v 1.2, 7 pages

We compute the Donaldson numbers $q_{17}(CP^2)=2540$ and $q_{21}(CP^2)=233208$., Comment: 8 pages. Submitted to the proceedings of the Dec. 94 Taniguchi Symposium "Vector Bundles". AMSLaTeX v 1.2 (latex2e, amsart, amscd)

We outline a strategy for computing intersection numbers on smooth varieties with torus actions using a residue formula of Bott. As an example, Gromov-Witten numbers of twisted cubic and elliptic quartic curves on some general complete intersection in projective space are computed. The results are consistent with predictions made from mirror symmetry computations. We also compute degrees of some loci in the linear system of plane curves of degrees less than 10, like those corresponding to sums of powers of linear forms, and curves carrying inscribed polygons., Comment: 22 pages, amslatex 1.1 The paper is a considerably expanded version of our previous eprint alg-geom/9409006 which had the title "Counting twisted cubics on general complete intersections"

This paper has been rendered obsolete by our newer eprint alg-geom/9411005 "Bott's formula and enumerative geometry", which is a considerably expanded version of the same paper, in spite of the change of titles. Please download alg-geom/9411005 instead, and update possible references., Comment: 14 pages, amslatex 1.1

"COM-MAND" represents the application of behavior modification concepts to the study, management, and control of behavior within a more traditional group therapy context. The emphasis of this system is on behavior, and is predicated on the learning/educational model. The therapist functions merely as an instructor, conducting a variety of behavioral training exercises. The client is taught to observe his own behavior, its consequences, and its effect upon others through the use of various prompts and instructions by the therapist/instructor, and he is directly and immediately reinforced when he does so. The system is presently being studied through use with inmate/addicts in a federal correctional institution. Details of the techniques used and hardware required, as well as an evaluation overview of the program are presented. (Author/PC)