-
A lattice framework for generalizing shellable complexes and matroids
Authors:
Rakhi Pratihar,
Tovohery H. Randrianarisoa,
Klara Stokes
Abstract:
We introduce the notion of power lattices that unifies and extends the equicardinal geometric lattices, Cartesian products of subspace lattices, and multiset subset lattices, among several others. The notions of shellability for simplicial complexes, q-complexes, and multicomplexes are then unified and extended to that of complexes in power lattices, which we name as P-complexes. A nontrivial clas…
▽ More
We introduce the notion of power lattices that unifies and extends the equicardinal geometric lattices, Cartesian products of subspace lattices, and multiset subset lattices, among several others. The notions of shellability for simplicial complexes, q-complexes, and multicomplexes are then unified and extended to that of complexes in power lattices, which we name as P-complexes. A nontrivial class of shellable P-complexes are obtained via P-complexes of the independent sets of a matroid in power lattice, which we introduce to generalize matroids in Boolean lattices, q-matroids in subspace lattices, and sum-matroids in Cartesian products of subspace lattices. We also prove that shellable P-complexes in a power lattice yield shellable order complexes, extending the celebrated result of shellability of order complexes of (equicardinal) geometric lattices by Björner and also, a recent result on shellability of order complexes of lexicographically shellable q-complexes. Finally, we provide a construction of matroids on the lattice of multiset subsets from weighted graphs. We also consider a variation of Stanley-Reisner rings associated with shellable multicomplexes than the one considered by Herzog and Popescu and proved that these rings are sequentially Cohen-Macaulay.
△ Less
Submitted 11 July, 2024;
originally announced July 2024.
-
Homotopy type of shellable $q$-complexes and their homology groups
Authors:
Sudhir R. Ghorpade,
Rakhi Pratihar,
Tovohery H. Randrianarisoa,
Hugues Verdure,
Glen Wilson
Abstract:
The theory of shellable simplicial complexes brings together combinatorics, algebra, and topology in a remarkable way. Initially introduced by Alder for $q$-simplicial complexes, recent work of Ghorpade, Pratihar, and Randrianarisoa extends the study of shellability to $q$-matroid complexes and determines singular homology groups for a subclass of these $q$-simplicial complexes. In this paper, we…
▽ More
The theory of shellable simplicial complexes brings together combinatorics, algebra, and topology in a remarkable way. Initially introduced by Alder for $q$-simplicial complexes, recent work of Ghorpade, Pratihar, and Randrianarisoa extends the study of shellability to $q$-matroid complexes and determines singular homology groups for a subclass of these $q$-simplicial complexes. In this paper, we determine the homotopy type of shellable $q$-simplicial complexes. Moreover, we establish the shellability of order complexes from lexicographically shellable $q$-simplicial complexes, that include the $q$-matroid complexes. This results in a comprehensive determination of the homology groups for any lexicographically shellable $q$-complexes.
△ Less
Submitted 11 March, 2024;
originally announced March 2024.
-
Deep Horseshoe Gaussian Processes
Authors:
Ismaël Castillo,
Thibault Randrianarisoa
Abstract:
Deep Gaussian processes have recently been proposed as natural objects to fit, similarly to deep neural networks, possibly complex features present in modern data samples, such as compositional structures. Adopting a Bayesian nonparametric approach, it is natural to use deep Gaussian processes as prior distributions, and use the corresponding posterior distributions for statistical inference. We i…
▽ More
Deep Gaussian processes have recently been proposed as natural objects to fit, similarly to deep neural networks, possibly complex features present in modern data samples, such as compositional structures. Adopting a Bayesian nonparametric approach, it is natural to use deep Gaussian processes as prior distributions, and use the corresponding posterior distributions for statistical inference. We introduce the deep Horseshoe Gaussian process Deep-HGP, a new simple prior based on deep Gaussian processes with a squared-exponential kernel, that in particular enables data-driven choices of the key lengthscale parameters. For nonparametric regression with random design, we show that the associated tempered posterior distribution recovers the unknown true regression curve optimally in terms of quadratic loss, up to a logarithmic factor, in an adaptive way. The convergence rates are simultaneously adaptive to both the smoothness of the regression function and to its structure in terms of compositions. The dependence of the rates in terms of dimension are explicit, allowing in particular for input spaces of dimension increasing with the number of observations.
△ Less
Submitted 4 March, 2024;
originally announced March 2024.
-
Variational Gaussian Processes For Linear Inverse Problems
Authors:
Thibault Randrianarisoa,
Botond Szabo
Abstract:
By now Bayesian methods are routinely used in practice for solving inverse problems. In inverse problems the parameter or signal of interest is observed only indirectly, as an image of a given map, and the observations are typically further corrupted with noise. Bayes offers a natural way to regularize these problems via the prior distribution and provides a probabilistic solution, quantifying the…
▽ More
By now Bayesian methods are routinely used in practice for solving inverse problems. In inverse problems the parameter or signal of interest is observed only indirectly, as an image of a given map, and the observations are typically further corrupted with noise. Bayes offers a natural way to regularize these problems via the prior distribution and provides a probabilistic solution, quantifying the remaining uncertainty in the problem. However, the computational costs of standard, sampling based Bayesian approaches can be overly large in such complex models. Therefore, in practice variational Bayes is becoming increasingly popular. Nevertheless, the theoretical understanding of these methods is still relatively limited, especially in context of inverse problems. In our analysis we investigate variational Bayesian methods for Gaussian process priors to solve linear inverse problems. We consider both mildly and severely ill-posed inverse problems and work with the popular inducing variables variational Bayes approach proposed by Titsias in 2009. We derive posterior contraction rates for the variational posterior in general settings and show that the minimax estimation rate can be attained by correctly tunned procedures. As specific examples we consider a collection of inverse problems including the heat equation, Volterra operator and Radon transform and inducing variable methods based on population and empirical spectral features.
△ Less
Submitted 1 November, 2023;
originally announced November 2023.
-
Sparse Incidence Geometries and Pebble Game Algorithms
Authors:
Signe Lundqvist,
Tovohery Randrianarisoa,
Klara Stokes,
Joannes Vermant
Abstract:
We generalize a sparsity condition for hypergraphs and show a result relating sparseness of hypergraphs to the decomposition of a modified incidence graph into edge-disjoint spanning forests. We also give new sparsity conditions for posets, and define an algorithm of pebble game type for recognising this sparsity.
We generalize a sparsity condition for hypergraphs and show a result relating sparseness of hypergraphs to the decomposition of a modified incidence graph into edge-disjoint spanning forests. We also give new sparsity conditions for posets, and define an algorithm of pebble game type for recognising this sparsity.
△ Less
Submitted 11 March, 2024; v1 submitted 8 June, 2023;
originally announced June 2023.
-
The Euler characteristic, $q$-matroids, and a Möbius function
Authors:
Trygve Johnsen,
Rakhi Pratihar,
Tovohery Hajatiana Randrianarisoa
Abstract:
We first give two new proofs of an old result that the reduced Euler characteristic of a matroid complex is equal to the Möbius number of the lattice of cycles of the matroid up to the sign. The purpose has been to find a model to establish an analogous result for the case of $q$-matroids and we find a relation between the Euler characteristic of the simplicial chain complex associated to a $q$-ma…
▽ More
We first give two new proofs of an old result that the reduced Euler characteristic of a matroid complex is equal to the Möbius number of the lattice of cycles of the matroid up to the sign. The purpose has been to find a model to establish an analogous result for the case of $q$-matroids and we find a relation between the Euler characteristic of the simplicial chain complex associated to a $q$-matroid complex and the lattice of $q$-cycles of the $q$-matroid. We use this formula to find the complete homology over $\mathbb{Z}$ of this shellable simplicial complex. We give a characterization of nonzero Euler characteristic for such order complexes. Finally, based on these results we remark why singular homology of a $q$-matroid equipped with order topology may not be effective to describe the $q$-cycles unlike the classical case of matroids.
△ Less
Submitted 22 October, 2022;
originally announced October 2022.
-
On Adaptive Confidence Sets for the Wasserstein Distances
Authors:
Neil Deo,
Thibault Randrianarisoa
Abstract:
In the density estimation model, we investigate the problem of constructing adaptive honest confidence sets with radius measured in Wasserstein distance $W_p$, $p\geq1$, and for densities with unknown regularity measured on a Besov scale. As sampling domains, we focus on the $d-$dimensional torus $\mathbb{T}^d$, in which case $1\leq p\leq 2$, and $\mathbb{R}^d$, for which $p=1$. We identify necess…
▽ More
In the density estimation model, we investigate the problem of constructing adaptive honest confidence sets with radius measured in Wasserstein distance $W_p$, $p\geq1$, and for densities with unknown regularity measured on a Besov scale. As sampling domains, we focus on the $d-$dimensional torus $\mathbb{T}^d$, in which case $1\leq p\leq 2$, and $\mathbb{R}^d$, for which $p=1$. We identify necessary and sufficient conditions for the existence of adaptive confidence sets with diameters of the order of the regularity-dependent $W_p$-minimax estimation rate. Interestingly, it appears that the possibility of such adaptation of the diameter depends on the dimension of the underlying space. In low dimensions, $d\leq 4$, adaptation to any regularity is possible. In higher dimensions, adaptation is possible if and only if the underlying regularities belong to some interval of width at least $d/(d-4)$. This contrasts with the usual $L_p-$theory where, independently of the dimension, adaptation requires regularities to lie in a small fixed-width window. For configurations allowing these adaptive sets to exist, we explicitly construct confidence regions via the method of risk estimation, centred at adaptive estimators. Those are the first results in a statistical approach to adaptive uncertainty quantification with Wasserstein distances. Our analysis and methods extend more globally to weak losses such as Sobolev norm distances with negative smoothness indices.
△ Less
Submitted 17 November, 2021; v1 submitted 16 November, 2021;
originally announced November 2021.
-
Optional Pólya trees: posterior rates and uncertainty quantification
Authors:
Ismaël Castillo,
Thibault Randrianarisoa
Abstract:
We consider statistical inference in the density estimation model using a tree-based Bayesian approach, with Optional Pólya trees as prior distribution. We derive near-optimal convergence rates for corresponding posterior distributions with respect to the supremum norm. For broad classes of Hölder-smooth densities, we show that the method automatically adapts to the unknown Hölder regularity param…
▽ More
We consider statistical inference in the density estimation model using a tree-based Bayesian approach, with Optional Pólya trees as prior distribution. We derive near-optimal convergence rates for corresponding posterior distributions with respect to the supremum norm. For broad classes of Hölder-smooth densities, we show that the method automatically adapts to the unknown Hölder regularity parameter. We consider the question of uncertainty quantification by providing mathematical guarantees for credible sets from the obtained posterior distributions, leading to near-optimal uncertainty quantification for the density function, as well as related functionals such as the cumulative distribution function. The results are illustrated through a brief simulation study.
△ Less
Submitted 11 October, 2021;
originally announced October 2021.
-
Shellability and homology of $q$-complexes and $q$-matroids
Authors:
Sudhir R. Ghorpade,
Rakhi Pratihar,
Tovohery H. Randrianarisoa
Abstract:
We consider a $q$-analogue of abstract simplicial complexes, called $q$-complexes, and discuss the notion of shellability for such complexes. It is shown that $q$-complexes formed by independent subspaces of a $q$-matroid are shellable. Further, we explicitly determine the homology of $q$-complexes corresponding to uniform $q$-matroids. We also outline some partial results concerning the determina…
▽ More
We consider a $q$-analogue of abstract simplicial complexes, called $q$-complexes, and discuss the notion of shellability for such complexes. It is shown that $q$-complexes formed by independent subspaces of a $q$-matroid are shellable. Further, we explicitly determine the homology of $q$-complexes corresponding to uniform $q$-matroids. We also outline some partial results concerning the determination of homology of arbitrary shellable $q$-complexes..
△ Less
Submitted 18 May, 2021; v1 submitted 25 February, 2021;
originally announced February 2021.
-
Smoothing and adaptation of shifted Pólya Tree ensembles
Authors:
Thibault Randrianarisoa
Abstract:
Recently, S. Arlot and R. Genuer have shown that a model of random forests outperforms its single-tree counterpart in the estimation of $α-$Hölder functions, $α\leq2$. This backs up the idea that ensembles of tree estimators are smoother estimators than single trees. On the other hand, most positive optimality results on Bayesian tree-based methods assume that $α\leq1$. Naturally, one wonders whet…
▽ More
Recently, S. Arlot and R. Genuer have shown that a model of random forests outperforms its single-tree counterpart in the estimation of $α-$Hölder functions, $α\leq2$. This backs up the idea that ensembles of tree estimators are smoother estimators than single trees. On the other hand, most positive optimality results on Bayesian tree-based methods assume that $α\leq1$. Naturally, one wonders whether Bayesian counterparts of forest estimators are optimal on smoother classes, just like it has been observed for frequentist estimators for $α\leq 2$. We dwell on the problem of density estimation and introduce an ensemble estimator from the classical (truncated) Pólya tree construction in Bayesian nonparametrics. The resulting Bayesian forest estimator is shown to lead to optimal posterior contraction rates, up to logarithmic terms, for the Hellinger and $L^1$ distances on probability density functions on $[0;1)$ for arbitrary Hölder regularity $α>0$. This improves upon previous results for constructions related to the Pólya tree prior whose optimality was only proven in the case $α\leq 1$. Also, we introduce an adaptive version of this new prior in the sense that it does not require the knowledge of $α$ to be defined and attain optimality.
△ Less
Submitted 15 September, 2021; v1 submitted 23 October, 2020;
originally announced October 2020.
-
Some Matroids Related to Sum-Rank Metric Codes
Authors:
Avijit Panja,
Rakhi Pratihar,
Tovohery Hajatiana Randrianarisoa
Abstract:
We introduce the notion of sum-matroids and show its association with sum-rank metric codes. As a consequence, some results for sum-rank metric codes by Martínez-Peñas are generalized for sum-matroids. The sum-matroids generalize the notions of matroids and $q$-matroids. We define the generalized weights for sum-matroids and prove a Wei-type duality theorem which generalizes the analogous cases fo…
▽ More
We introduce the notion of sum-matroids and show its association with sum-rank metric codes. As a consequence, some results for sum-rank metric codes by Martínez-Peñas are generalized for sum-matroids. The sum-matroids generalize the notions of matroids and $q$-matroids. We define the generalized weights for sum-matroids and prove a Wei-type duality theorem which generalizes the analogous cases for matroids and q-matroids.
△ Less
Submitted 22 March, 2022; v1 submitted 20 December, 2019;
originally announced December 2019.
-
The number of matrices over $\mathbb{F}_q$ with irreducible characteristic polynomial
Authors:
Tovohery Hajatiana Randrianarisoa
Abstract:
Let $\mathbb{F}_q$ be a finite field with $q$ elements. M. Gerstenhaber and Irving Reiner has given two different methods to show the number of matrices with a given characteristic polynomial. In this talk, we will give another proof for the particular case where the characteristic polynomial is irreducible. The number of such matrices is important to know the efficiency of an algorithm to factor…
▽ More
Let $\mathbb{F}_q$ be a finite field with $q$ elements. M. Gerstenhaber and Irving Reiner has given two different methods to show the number of matrices with a given characteristic polynomial. In this talk, we will give another proof for the particular case where the characteristic polynomial is irreducible. The number of such matrices is important to know the efficiency of an algorithm to factor polynomials using Drinfeld modules.
△ Less
Submitted 12 February, 2014;
originally announced February 2014.
