arXiv subjects

Nicolas Bousquet

Publications and source records attributed to Nicolas Bousquet.

At least 55 records · Page 3Linked to original sources

arXiv2023

A note on the flip distance between non-crossing spanning trees

We consider spanning trees of $n$ points in convex position whose edges are pairwise non-crossing. Applying a flip to such a tree consists in adding an edge and removing another so that the result is still a non-crossing spanning tree. Given two trees, we investigate the minimum number of flips required to transform one into the other. The naive $2n-Ω(1)$ upper bound stood for 25 years until a recent breakthrough from Aichholzer et al. yielding a $2n-Ω(\log n)$ bound. We improve their result with a $2n-Ω(\sqrt{n})$ upper bound, and we strengthen and shorten the proofs of several of their results.

cs.CG↗

arXiv2023

On the coalitional decomposition of parameters of interest

Understanding the behavior of a black-box model with probabilistic inputs can be based on the decomposition of a parameter of interest (e.g., its variance) into contributions attributed to each coalition of inputs (i.e., subsets of inputs). In this paper, we produce conditions for obtaining unambiguous and interpretable decompositions of very general parameters of interest. This allows to recover known decompositions, holding under weaker assumptions than stated in the literature.

math.ST↗

arXiv2023

Extremal Independent Set Reconfiguration

The independent set reconfiguration problem asks whether one can transform one given independent set of a graph into another, by changing vertices one by one in such a way the intermediate sets remain independent. Extremal problems on independent sets are widely studied: for example, it is well known that an $n$-vertex graph has at most $3^{n/3}$ maximum independent sets (and this is tight). This paper investigates the asymptotic behavior of maximum possible length of a shortest reconfiguration sequence for independent sets of size $k$ among all $n$-vertex graphs. We give a tight bound for $k=2$. We also provide a subquadratic upper bound (using the hypergraph removal lemma) as well as an almost tight construction for $k=3$. We generalize our results for larger values of $k$ by proving an $n^{2\lfloor k/3 \rfloor}$ lower bound.

math.CO↗

arXiv2022

Quantile-constrained Wasserstein projections for robust interpretability of numerical and machine learning models

Robustness studies of black-box models is recognized as a necessary task for numerical models based on structural equations and predictive models learned from data. These studies must assess the model's robustness to possible misspecification of regarding its inputs (e.g., covariate shift). The study of black-box models, through the prism of uncertainty quantification (UQ), is often based on sensitivity analysis involving a probabilistic structure imposed on the inputs, while ML models are solely constructed from observed data. Our work aim at unifying the UQ and ML interpretability approaches, by providing relevant and easy-to-use tools for both paradigms. To provide a generic and understandable framework for robustness studies, we define perturbations of input information relying on quantile constraints and projections with respect to the Wasserstein distance between probability measures, while preserving their dependence structure. We show that this perturbation problem can be analytically solved. Ensuring regularity constraints by means of isotonic polynomial approximations leads to smoother perturbations, which can be more suitable in practice. Numerical experiments on real case studies, from the UQ and ML fields, highlight the computational feasibility of such studies and provide local and global insights on the robustness of black-box models to input perturbations.

math.OC↗

arXiv2022

Locating-dominating sets: from graphs to oriented graphs

A locating-dominating set in an undirected graph is a subset of vertices $S$ such that $S$ is dominating and for every $u,v \notin S$, we have $N(u)\cap S\ne N(v)\cap S$. In this paper, we consider the oriented version of the problem. A locating-dominating set in an oriented graph is a set $S$ such that for every $w\in V$, $N[w]^-\cap S=\emptyset$ and for each pair of vertices $u,v\in V\setminus S$, $N^-(u)\cap S\ne N^-(v)\cap S$. We consider the following two parameters. Given an undirected graph $G$, we look for $\overset{\rightarrow}γ_{LD}(G)$ ($\overset{\rightarrow}Γ_{LD}(G))$ which is the size of the smallest (largest) optimal locating-dominating set over all orientations of $G$. In particular, if $D$ is an orientation of $G$, then $\overset{\rightarrow}γ_{LD}(G)\leqγ_{LD}(D)\leq\overset{\rightarrow}Γ_{LD}(G)$. For the best orientation, we prove that, for every twin-free graph $G$ on $n$ vertices, $\overset{\rightarrow}γ_{LD}(G)\le n/2$ proving a ``directed version'' of a conjecture on $γ_{LD}(G)$. Moreover, we give some bounds for $\overset{\rightarrow}γ_{LD}(G)$ on many graph classes and drastically improve the value $n/2$ for (almost) $d$-regular graphs by showing that $\overset{\rightarrow}γ_{LD}(G)\in O(\log d/d\cdot n)$ using a probabilistic argument. While $\overset{\rightarrow}γ_{LD}(G)\leqγ_{LD}(G)$ holds for every graph $G$, we give some graph classes graphs for which $\overset{\rightarrow}Γ_{LD}(G)\geqγ_{LD}(G)$ and some for which $\overset{\rightarrow}Γ_{LD}(G)\leq γ_{LD}(G)$. We also give general bounds for $\overset{\rightarrow}Γ_{LD}(G)$. Finally, we show that for many graph classes $\overset{\rightarrow}Γ_{LD}(G)$ is polynomial on $n$ but we leave open the question whether there exist graphs with $\overset{\rightarrow}Γ_{LD}(G)\in O(\log n)$.

math.CO↗

arXiv2022

Metric dimension on sparse graphs and its applications to zero forcing sets

The metric dimension dim(G) of a graph $G$ is the minimum cardinality of a subset $S$ of vertices of $G$ such that each vertex of $G$ is uniquely determined by its distances to $S$. It is well-known that the metric dimension of a graph can be drastically increased by the modification of a single edge. Our main result consists in proving that the increase of the metric dimension of an edge addition can be amortized in the sense that if the graph consists of a spanning tree $T$ plus $c$ edges, then the metric dimension of $G$ is at most the metric dimension of $T$ plus $6c$. We then use this result to prove a weakening of a conjecture of Eroh et al. The zero forcing number $Z(G)$ of $G$ is the minimum cardinality of a subset $S$ of black vertices (whereas the other vertices are colored white) of $G$ such that all the vertices will turned black after applying finitely many times the following rule: a white vertex is turned black if it is the only white neighbor of a black vertex. Eroh et al. conjectured that, for any graph $G$, $dim(G)\leq Z(G) + c(G)$, where $c(G)$ is the number of edges that have to be removed from $G$ to get a forest. They proved the conjecture is true for trees and unicyclic graphs. We prove a weaker version of the conjecture: $dim(G)\leq Z(G)+6c(G)$ holds for any graph. We also prove that the conjecture is true for graphs with edge disjoint cycles, widely generalizing the unicyclic result of Eroh et al.

math.CO↗

arXiv2022

Token sliding on graphs of girth five

In the Token Sliding problem we are given a graph $G$ and two independent sets $I_s$ and $I_t$ in $G$ of size $k \geq 1$. The goal is to decide whether there exists a sequence $\langle I_1, I_2, \ldots, I_\ell \rangle$ of independent sets such that for all $i \in \{1,\ldots, \ell\}$ the set $I_i$ is an independent set of size $k$, $I_1 = I_s$, $I_\ell = I_t$ and $I_i \triangle I_{i + 1} = \{u, v\} \in E(G)$. Intuitively, we view each independent set as a collection of tokens placed on the vertices of the graph. Then, the problem asks whether there exists a sequence of independent sets that transforms $I_s$ into $I_t$ where at each step we are allowed to slide one token from a vertex to a neighboring vertex. In this paper, we focus on the parameterized complexity of Token Sliding parameterized by $k$. As shown by Bartier et al., the problem is W[1]-hard on graphs of girth four or less, and the authors posed the question of whether there exists a constant $p \geq 5$ such that the problem becomes fixed-parameter tractable on graphs of girth at least $p$. We answer their question positively and prove that the problem is indeed fixed-parameter tractable on graphs of girth five or more, which establishes a full classification of the tractability of Token Sliding parameterized by the number of tokens based on the girth of the input graph.

cs.CC↗

arXiv2022

A survey on the parameterized complexity of the independent set and (connected) dominating set reconfiguration problems

A graph vertex-subset problem defines which subsets of the vertices of an input graph are feasible solutions. We view a feasible solution as a set of tokens placed on the vertices of the graph. A reconfiguration variant of a vertex-subset problem asks, given two feasible solutions of size $k$, whether it is possible to transform one into the other by a sequence of token slides (along edges of the graph) or token jumps (between arbitrary vertices of the graph) such that each intermediate set remains a feasible solution of size $k$. Many algorithmic questions present themselves in the form of reconfiguration problems: Given the description of an initial system state and the description of a target state, is it possible to transform the system from its initial state into the target one while preserving certain properties of the system in the process? Such questions have received a substantial amount of attention under the so-called combinatorial reconfiguration framework. We consider reconfiguration variants of three fundamental underlying graph vertex-subset problems, namely Independent Set, Dominating Set, and Connected Dominating Set. We survey both older and more recent work on the parameterized complexity of all three problems when parameterized by the number of tokens $k$. The emphasis will be on positive results and the most common techniques for the design of fixed-parameter tractable algorithms.

cs.CC↗

arXiv2022

Galactic Token Sliding

Given a graph $G$ and two independent sets $I_s$ and $I_t$ of size $k$, the independent set reconfiguration problem asks whether there exists a sequence of $k$-sized independent sets $I_s = I_0, I_1, I_2, \ldots, I_\ell = I_t$ such that each independent set is obtained from the previous one using a so-called reconfiguration step. Viewing each independent set as a collection of $k$ tokens placed on the vertices of a graph $G$, the two most studied reconfiguration steps are token jumping and token sliding. In the token jumping variant of the problem, a single step allows a token to jump from one vertex to any other vertex in the graph. In the token sliding variant, a token is only allowed to slide from a vertex to one of its neighbors. Like the independent set problem, both of the aforementioned problems are known to be W[1]-hard on general graphs. A very fruitful line of research has showed that the independent set problem becomes fixed-parameter tractable when restricted to sparse graph classes, such as planar, bounded treewidth, nowhere-dense, and all the way to biclique-free graphs. Over a series of papers, the same was shown to hold for the token jumping problem. As for the token sliding problem, which is mentioned in most of these papers, almost nothing is known beyond the fact that the problem is polynomial-time solvable on trees and interval graphs. We remedy this situation by introducing a new model for the reconfiguration of independent sets, which we call galactic reconfiguration. Using this new model, we show that (standard) token sliding is fixed-parameter tractable on graphs of bounded degree, planar graphs, and chordal graphs of bounded clique number. We believe that the galactic reconfiguration model is of independent interest and could potentially help in resolving the remaining open questions concerning the (parameterized) complexity of token sliding.

cs.CC↗

arXiv2022

Square coloring planar graphs with automatic discharging

The discharging method is a powerful proof technique, especially for graph coloring problems. Its major downside is that it often requires lengthy case analyses, which are sometimes given to a computer for verification. However, it is much less common to use a computer to actively look for a discharging proof. In this paper, we use a Linear Programming approach to automatically look for a discharging proof. While our system is not entirely autonomous, we manage to make some progress towards Wegner's conjecture for distance-$2$ coloring of planar graphs, by showing that $12$ colors are sufficient to color at distance $2$ every planar graph with maximum degree $4$.

math.CO↗

arXiv2022

What can be certified compactly?

Local certification consists in assigning labels (called \emph{certificates}) to the nodes of a network to certify a property of the network or the correctness of a data structure distributed on the network. The verification of this certification must be local: a node typically sees only its neighbors in the network. The main measure of performance of a certification is the size of its certificates. In 2011, Göös and Suomela identified $Θ(\log n)$ as a special certificate size: below this threshold little is possible, and several key properties do have certifications of this type. A certification with such small certificates is now called a \emph{compact local certification}, and it has become the gold standard of the area, similarly to polynomial time for centralized computing. A major question is then to understand which properties have $O(\log n)$ certificates, or in other words: what is the power of compact local certification? Recently, a series of papers have proved that several well-known network properties have compact local certifications: planarity, bounded-genus, etc. But one would like to have more general results, \emph{i.e.} meta-theorems. In the analogue setting of polynomial-time centralized algorithms, a very fruitful approach has been to prove that restricted types of problems can be solved in polynomial time in graphs with restricted structures. These problems are typically those that can be expressed in some logic, and the graph structures are whose with bounded width or depth parameters. We take a similar approach and prove the first meta-theorems for local certification. (See the abstract of the pdf for more details.)

cs.DC↗

arXiv2022

Chordal directed graphs are not $χ$-bounded

We show that digraphs with no transitive tournament on $3$ vertices and in which every induced directed cycle has length $3$ can have arbitrarily large dichromatic number. This answers to the negative a question of Carbonero, Hompe, Moore, and Spirkl (and extends some of their results).

math.CO↗

arXiv2022

Reconfiguration of Spanning Trees with Degree Constraint or Diameter Constraint

We investigate the complexity of finding a transformation from a given spanning tree in a graph to another given spanning tree in the same graph via a sequence of edge flips. The exchange property of the matroid bases immediately yields that such a transformation always exists if we have no constraints on spanning trees. In this paper, we wish to find a transformation which passes through only spanning trees satisfying some constraint. Our focus is bounding either the maximum degree or the diameter of spanning trees, and we give the following results. The problem with a lower bound on maximum degree is solvable in polynomial time, while the problem with an upper bound on maximum degree is PSPACE-complete. The problem with a lower bound on diameter is NP-hard, while the problem with an upper bound on diameter is solvable in polynomial time.

cs.DS↗

arXiv2021

Improved square coloring of planar graphs

Square coloring is a variant of graph coloring where vertices within distance two must receive different colors. When considering planar graphs, the most famous conjecture (Wegner, 1977) states that $\frac32Δ+1$ colors are sufficient to square color every planar graph of maximum degree $Δ$. This conjecture has been proven asymptotically for graphs with large maximum degree. We consider here planar graphs with small maximum degree and show that $2Δ+7$ colors are sufficient, which improves the best known bounds when $6\leqslant Δ\leqslant 31$.

math.CO↗

arXiv2021

Recolouring planar graphs of girth at least five

For a positive integer $k$, the $k$-recolouring graph of a graph $G$ has as vertex set all proper $k$-colourings of $G$ with two $k$-colourings being adjacent if they differ by the colour of exactly one vertex. A result of Dyer et al. regarding graphs of bounded degeneracy implies that the $7$-recolouring graphs of planar graphs, the $5$-recolouring graphs of triangle-free planar graphs and the $4$-recolouring graphs planar graphs of girth at least six are connected. On the other hand, there are planar graphs whose $6$-recolouring graph is disconnected, triangle-free planar graphs whose $4$-recolouring graph is disconnected and planar graphs of any given girth whose $3$-recolouring graph is disconnected. The main result of this paper consists in showing, via a novel application of the discharging method, that the $4$-recolouring graph of every planar graph of girth five is connected. This completes the classification of the connectedness of the recolouring graph for planar graphs of given girth. We also prove some theorems regarding the diameter of the recolouring graph of planar graphs.

math.CO↗

arXiv2021

EPTAS and Subexponential Algorithm for Maximum Clique on Disk and Unit Ball Graphs

A (unit) disk graph is the intersection graph of closed (unit) disks in the plane. Almost three decades ago, an elegant polynomial-time algorithm was found for \textsc{Maximum Clique} on unit disk graphs [Clark, Colbourn, Johnson; Discrete Mathematics '90]. Since then, it has been an intriguing open question whether or not tractability can be extended to general disk graphs. We show that the disjoint union of two odd cycles is never the complement of a disk graph nor of a unit (3-dimensional) ball graph. From that fact and existing results, we derive a simple QPTAS and a subexponential algorithm running in time $2^{\tilde{O}(n^{2/3})}$ for \textsc{Maximum Clique} on disk and unit ball graphs. We then obtain a randomized EPTAS for computing the independence number on graphs having no disjoint union of two odd cycles as an induced subgraph, bounded VC-dimension, and linear independence number. This, in combination with our structural results, yields a randomized EPTAS for \textsc{Max Clique} on disk and unit ball graphs. \textsc{Max Clique} on unit ball graphs is equivalent to finding, given a collection of points in $\mathbb R^3$, a maximum subset of points with diameter at most some fixed value. In stark contrast, \textsc{Maximum Clique} on ball graphs and unit $4$-dimensional ball graphs, as well as intersection graphs of filled ellipses (even close to unit disks) or filled triangles is unlikely to have such algorithms. Indeed, we show that, for all those problems, there is a constant ratio of approximation which cannot be attained even in time $2^{n^{1-\varepsilon}}$, unless the Exponential Time Hypothesis fails.

cs.DS↗

arXiv2021

Degeneracy of $P_t$-free and $C_{\geq t}$-free graphs with no large complete bipartite subgraphs

A hereditary class of graphs $\mathcal{G}$ is \emph{$χ$-bounded} if there exists a function $f$ such that every graph $G \in \mathcal{G}$ satisfies $χ(G) \leq f(ω(G))$, where $χ(G)$ and $ω(G)$ are the chromatic number and the clique number of $G$, respectively. As one of the first results about $χ$-bounded classes, Gyárfás proved in 1985 that if $G$ is $P_t$-free, i.e., does not contain a $t$-vertex path as an induced subgraph, then $χ(G) \leq (t-1)^{ω(G)-1}$. In 2017, Chudnovsky, Scott, and Seymour proved that $C_{\geq t}$-free graphs, i.e., graphs that exclude induced cycles with at least $t$ vertices, are $χ$-bounded as well, and the obtained bound is again superpolynomial in the clique number. Note that $P_{t-1}$-free graphs are in particular $C_{\geq t}$-free. It remains a major open problem in the area whether for $C_{\geq t}$-free, or at least $P_t$-free graphs $G$, the value of $χ(G)$ can be bounded from above by a polynomial function of $ω(G)$. We consider a relaxation of this problem, where we compare the chromatic number with the size of a largest balanced biclique contained in the graph as a (not necessarily induced) subgraph. We show that for every $t$ there exists a constant $c$ such that for and every $C_{\geq t}$-free graph which does not contain $K_{\ell,\ell}$ as a subgraph, it holds that $χ(G) \leq \ell^{c}$.

math.CO↗

arXiv2021

Local certification of MSO properties for bounded treedepth graphs

The graph model checking problem consists in testing whether an input graph satisfies a given logical formula. In this paper, we study this problem in a distributed setting, namely local certification. The goal is to assign labels to the nodes of a network to certify that some given property is satisfied, in such a way that the labels can be checked locally. We first investigate which properties can be locally certified with small certificates. Not surprisingly, this is almost never the case, except for not very expressive logic fragments. Following the steps of Courcelle-Grohe, we then look for meta-theorems explaining what happens when we parameterize the problem by some standard measures of how simple the graph classes are. In that direction, our main result states that any MSO formula can be locally certified on graphs with bounded treedepth with a logarithmic number of bits per node, which is the golden standard in certification.

cs.DC↗