arXiv subjects

Anand Louis

Publications and source records attributed to Anand Louis.

At least 37 records · Page 2Linked to original sources

arXiv2021

Group Fairness for Knapsack Problems

We study the knapsack problem with group fairness constraints. The input of the problem consists of a knapsack of bounded capacity and a set of items, each item belongs to a particular category and has and associated weight and value. The goal of this problem is to select a subset of items such that all categories are fairly represented, the total weight of the selected items does not exceed the capacity of the knapsack,and the total value is maximized. We study the fairness parameters such as the bounds on the total value of items from each category, the total weight of items from each category, and the total number of items from each category. We give approximation algorithms for these problems. These fairness notions could also be extended to the min-knapsack problem. The fair knapsack problems encompass various important problems, such as participatory budgeting, fair budget allocation, advertising.

cs.DS↗

arXiv2020

Planted Models for the Densest $k$-Subgraph Problem

Given an undirected graph $G$, the Densest $k$-subgraph problem (DkS) asks to compute a set $S \subset V$ of cardinality $\left\lvert S\right\rvert \leq k$ such that the weight of edges inside $S$ is maximized. This is a fundamental NP-hard problem whose approximability, inspite of many decades of research, is yet to be settled. The current best known approximation algorithm due to Bhaskara et al. (2010) computes a $\mathcal{O}\left({n^{1/4 + ε}}\right)$ approximation in time $n^{\mathcal{O}\left(1/ε\right)}$, for any $ε> 0$. We ask what are some "easier" instances of this problem? We propose some natural semi-random models of instances with a planted dense subgraph, and study approximation algorithms for computing the densest subgraph in them. These models are inspired by the semi-random models of instances studied for various other graph problems such as the independent set problem, graph partitioning problems etc. For a large range of parameters of these models, we get significantly better approximation factors for the Densest $k$-subgraph problem. Moreover, our algorithm recovers a large part of the planted solution.

cs.DS↗

arXiv2020

Stability of Linear Structural Equation Models of Causal Inference

We consider the numerical stability of the parameter recovery problem in Linear Structural Equation Model ($\LSEM$) of causal inference. A long line of work starting from Wright (1920) has focused on understanding which sub-classes of $\LSEM$ allow for efficient parameter recovery. Despite decades of study, this question is not yet fully resolved. The goal of this paper is complementary to this line of work; we want to understand the stability of the recovery problem in the cases when efficient recovery is possible. Numerical stability of Pearl's notion of causality was first studied in Schulman and Srivastava (2016) using the concept of condition number where they provide ill-conditioned examples. In this work, we provide a condition number analysis for the $\LSEM$. First we prove that under a sufficient condition, for a certain sub-class of $\LSEM$ that are \emph{bow-free} (Brito and Pearl (2002)), the parameter recovery is stable. We further prove that \emph{randomly} chosen input parameters for this family satisfy the condition with a substantial probability. Hence for this family, on a large subset of parameter space, recovery is numerically stable. Next we construct an example of $\LSEM$ on four vertices with \emph{unbounded} condition number. We then corroborate our theoretical findings via simulations as well as real-world experiments for a sociology application. Finally, we provide a general heuristic for estimating the condition number of any $\LSEM$ instance.

cs.LG↗

arXiv2020

Robust Identifiability in Linear Structural Equation Models of Causal Inference

In this work, we consider the problem of robust parameter estimation from observational data in the context of linear structural equation models (LSEMs). LSEMs are a popular and well-studied class of models for inferring causality in the natural and social sciences. One of the main problems related to LSEMs is to recover the model parameters from the observational data. Under various conditions on LSEMs and the model parameters the prior work provides efficient algorithms to recover the parameters. However, these results are often about generic identifiability. In practice, generic identifiability is not sufficient and we need robust identifiability: small changes in the observational data should not affect the parameters by a huge amount. Robust identifiability has received far less attention and remains poorly understood. Sankararaman et al. (2019) recently provided a set of sufficient conditions on parameters under which robust identifiability is feasible. However, a limitation of their work is that their results only apply to a small sub-class of LSEMs, called ``bow-free paths.'' In this work, we significantly extend their work along multiple dimensions. First, for a large and well-studied class of LSEMs, namely ``bow free'' models, we provide a sufficient condition on model parameters under which robust identifiability holds, thereby removing the restriction of paths required by prior work. We then show that this sufficient condition holds with high probability which implies that for a large set of parameters robust identifiability holds and that for such parameters, existing algorithms already achieve robust identifiability. Finally, we validate our results on both simulated and real-world datasets.

cs.LG↗

arXiv2020

Approximation Algorithms and Hardness for Strong Unique Games

The UNIQUE GAMES problem is a central problem in algorithms and complexity theory. Given an instance of UNIQUE GAMES, the STRONG UNIQUE GAMES problem asks to find the largest subset of vertices, such that the UNIQUE GAMES instance induced on them is completely satisfiable. In this paper, we give new algorithmic and hardness results for STRONG UNIQUE GAMES. Given an instance with label set size $k$ where a set of $(1 - ε)$-fraction of the vertices induce an instance that is completely satisfiable, our first algorithm produces a set of $1 - \widetilde{O}({k^2}) ε\sqrt{\log n}$ fraction of the vertices such that the UNIQUE GAMES induced on them is completely satisfiable. In the same setting, our second algorithm produces a set of $1 - \widetilde{O}({k^2}) \sqrt{ε\log d}$ (here $d$ is the largest vertex degree of the graph) fraction of the vertices such that the UNIQUE GAMES induced on them is completely satisfiable. The technical core of our results is a new connection between STRONG UNIQUE GAMES and Small-Set-Vertex-Expansion in graphs. Complementing this, assuming the Unique Games Conjecture, we prove that it is NP-hard to compute a set of size larger than $1 - Ω( \sqrt{ε\log k \log d})$ for which all the constraints induced on this set are satisfied. Given an undirected graph $G(V,E)$ the ODD CYCLE TRANSVERSAL problem asks to delete the least fraction of vertices to make the induced graph on the remaining vertices bipartite. As a corollary to our main algorithmic results, we obtain an algorithm that outputs a set $S$ such the graph induced on $V \setminus S$ is bipartite, and $|S|/n \leq O(\sqrt{ε\log d})$ (here $d$ is the largest vertex degree and $ε$ is the optimal fraction of vertices that need to be deleted). Assuming the Unique Games Conjecture, we prove a matching (up to constant factors) hardness.

cs.DS↗

arXiv2019

Planted Models for $k$-way Edge and Vertex Expansion

Graph partitioning problems are a central topic of study in algorithms and complexity theory. Edge expansion and vertex expansion, two popular graph partitioning objectives, seek a $2$-partition of the vertex set of the graph that minimizes the considered objective. However, for many natural applications, one might require a graph to be partitioned into $k$ parts, for some $k \geq 2$. For a $k$-partition $S_1, \ldots, S_k$ of the vertex set of a graph $G = (V,E)$, the $k$-way edge expansion (resp. vertex expansion) of $\{S_1, \ldots, S_k\}$ is defined as $\max_{i \in [k]} Φ(S_i)$, and the balanced $k$-way edge expansion (resp. vertex expansion) of $G$ is defined as \[ \min_{ \{S_1, \ldots, S_k\} \in \mathcal{P}_k} \max_{i \in [k]} Φ(S_i) \, , \] where $\mathcal{P}_k$ is the set of all balanced $k$-partitions of $V$ (i.e each part of a $k$-partition in $\mathcal{P}_k$ should have cardinality $|V|/k$), and $Φ(S)$ denotes the edge expansion (resp. vertex expansion) of $S \subset V$. We study a natural planted model for graphs where the vertex set of a graph has a $k$-partition $S_1, \ldots, S_k$ such that the graph induced on each $S_i$ has large expansion, but each $S_i$ has small edge expansion (resp. vertex expansion) in the graph. We give bi-criteria approximation algorithms for computing the balanced $k$-way edge expansion (resp. vertex expansion) of instances in this planted model.

cs.DS↗

arXiv2019

Approximation Algorithms for Partially Colorable Graphs

Graph coloring problems are a central topic of study in the theory of algorithms. We study the problem of partially coloring partially colorable graphs. For $α\leq 1$ and $k \in \mathbb{Z}^+$, we say that a graph $G=(V,E)$ is $α$-partially $k$-colorable, if there exists a subset $S\subset V$ of cardinality $ |S | \geq α| V |$ such that the graph induced on $S$ is $k$-colorable. Partial $k$-colorability is a more robust structural property of a graph than $k$-colorability. For graphs that arise in practice, partial $k$-colorability might be a better notion to use than $k$-colorability, since data arising in practice often contains various forms of noise. We give a polynomial time algorithm that takes as input a $(1 - ε)$-partially $3$-colorable graph $G$ and a constant $γ\in [ε, 1/10]$, and colors a $(1 - ε/γ)$ fraction of the vertices using $\tilde{O}\left(n^{0.25 + O(γ^{1/2})} \right)$ colors. We also study natural semi-random families of instances of partially $3$-colorable graphs and partially $2$-colorable graphs, and give stronger bi-criteria approximation guarantees for these family of instances.

cs.DS↗

arXiv2019

HyperGCN: A New Method of Training Graph Convolutional Networks on Hypergraphs

In many real-world network datasets such as co-authorship, co-citation, email communication, etc., relationships are complex and go beyond pairwise. Hypergraphs provide a flexible and natural modeling tool to model such complex relationships. The obvious existence of such complex relationships in many real-world networks naturaly motivates the problem of learning with hypergraphs. A popular learning paradigm is hypergraph-based semi-supervised learning (SSL) where the goal is to assign labels to initially unlabeled vertices in a hypergraph. Motivated by the fact that a graph convolutional network (GCN) has been effective for graph-based SSL, we propose HyperGCN, a novel GCN for SSL on attributed hypergraphs. Additionally, we show how HyperGCN can be used as a learning-based approach for combinatorial optimisation on NP-hard hypergraph problems. We demonstrate HyperGCN's effectiveness through detailed experimentation on real-world hypergraphs.

cs.LG↗

arXiv2019

On Euclidean $k$-Means Clustering with $α$-Center Proximity

$k$-means clustering is NP-hard in the worst case but previous work has shown efficient algorithms assuming the optimal $k$-means clusters are \emph{stable} under additive or multiplicative perturbation of data. This has two caveats. First, we do not know how to efficiently verify this property of optimal solutions that are NP-hard to compute in the first place. Second, the stability assumptions required for polynomial time $k$-means algorithms are often unreasonable when compared to the ground-truth clusters in real-world data. A consequence of multiplicative perturbation resilience is \emph{center proximity}, that is, every point is closer to the center of its own cluster than the center of any other cluster, by some multiplicative factor $α> 1$. We study the problem of minimizing the Euclidean $k$-means objective only over clusterings that satisfy $α$-center proximity. We give a simple algorithm to find the optimal $α$-center-proximal $k$-means clustering in running time exponential in $k$ and $1/(α- 1)$ but linear in the number of points and the dimension. We define an analogous $α$-center proximity condition for outliers, and give similar algorithmic guarantees for $k$-means with outliers and $α$-center proximity. On the hardness side we show that for any $α' > 1$, there exists an $α\leq α'$, $(α>1)$, and an $\varepsilon_0 > 0$ such that minimizing the $k$-means objective over clusterings that satisfy $α$-center proximity is NP-hard to approximate within a multiplicative $(1+\varepsilon_0)$ factor.

cs.DS↗

arXiv2018

Semi-Random Graphs with Planted Sparse Vertex Cuts: Algorithms for Exact and Approximate Recovery

The problem of computing the vertex expansion of a graph is an NP-hard problem. The current best worst-case approximation guarantees for computing the vertex expansion of a graph are a $O(\sqrt{\log n})$-approximation algorithm due to Feige, Hajiaghayi and Lee [SIAM J. Comp., 2008], and $O(\sqrt{OPT \log d})$ bound in graphs having vertex degrees at most $d$, due to Louis, Raghavendra and Vempala [FOCS 2013]. We study a natural semi-random model of graphs with sparse vertex cuts. For certain ranges of parameters, we give an algorithm to recover the planted sparse vertex cut exactly. For a larger range of parameters, we give a constant factor bi-criteria approximation algorithm to compute the graph's balanced vertex expansion. Our algorithms are based on studying a semidefinite programming relaxation for the balanced vertex expansion of the graph. In addition to being a family of instances that will help us to better understand the complexity of the computation of vertex expansion, our model can also be used in the study of community detection where only a few nodes from each community interact with nodes from other communities. There has been a lot of work on studying random and semi-random graphs with planted sparse edge cuts. To the best of our knowledge, our model of semi-random graphs with planted sparse vertex cuts has not been studied before.

cs.DS↗

arXiv2016

Spectral Properties of Hypergraph Laplacian and Approximation Algorithms

The celebrated Cheeger's Inequality establishes a bound on the edge expansion of a graph via its spectrum. This inequality is central to a rich spectral theory of graphs, based on studying the eigenvalues and eigenvectors of the adjacency matrix (and other related matrices) of graphs. It has remained open to define a suitable spectral model for hypergraphs whose spectra can be used to estimate various combinatorial properties of the hypergraph. In this paper we introduce a new hypergraph Laplacian operator generalizing the Laplacian matrix of graphs. In particular, the operator is induced by a diffusion process on the hypergraph, such that within each hyperedge, measure flows from vertices having maximum weighted measure to those having minimum. Since the operator is non-linear, we have to exploit other properties of the diffusion process to recover a spectral property concerning the "second eigenvalue" of the resulting Laplacian. Moreover, we show that higher order spectral properties cannot hold in general using the current framework. We consider a stochastic diffusion process, in which each vertex also experiences Brownian noise from outside the system. We show a relationship between the second eigenvalue and the convergence behavior of the process. We show that various hypergraph parameters like multi-way expansion and diameter can be bounded using this operator's spectral properties. Since higher order spectral properties do not hold for the Laplacian operator, we instead use the concept of procedural minimizers to consider higher order Cheeger-like inequalities. For any positive integer $k$, we give a polynomial time algorithm to compute an $O(\log r)$-approximation to the $k$-th procedural minimizer, where $r$ is the maximum cardinality of a hyperedge. We show that this approximation factor is optimal under the SSE hypothesis for constant values of $k$.

cs.DM↗

arXiv2016

Accelerated Newton Iteration: Roots of Black Box Polynomials and Matrix Eigenvalues

We study the problem of computing the largest root of a real rooted polynomial $p(x)$ to within error $\varepsilon $ given only black box access to it, i.e., for any $x \in {\mathbb R}$, the algorithm can query an oracle for the value of $p(x)$, but the algorithm is not allowed access to the coefficients of $p(x)$. A folklore result for this problem is that the largest root of a polynomial can be computed in $O(n \log (1/\varepsilon ))$ polynomial queries using the Newton iteration. We give a simple algorithm that queries the oracle at only $O(\log n \log(1/\varepsilon ))$ points, where $n$ is the degree of the polynomial. Our algorithm is based on a novel approach for accelerating the Newton method by using higher derivatives. As a special case, we consider the problem of computing the top eigenvalue of a symmetric matrix in ${\mathbb Q}^{n \times n}$ to within error $\varepsilon $ in time polynomial in the input description, i.e., the number of bits to describe the matrix and $\log(1/\varepsilon )$. Well-known methods such as the power iteration and Lanczos iteration incur running time polynomial in $1/\varepsilon $, while Gaussian elimination takes $Ω(n^4)$ bit operations. As a corollary of our main result, we obtain a $\tilde{O}(n^ω \log^2 ( ||A||_F/\varepsilon ))$ bit complexity algorithm to compute the top eigenvalue of the matrix $A$ or to check if it is approximately PSD ($A \succeq -\varepsilon I$).

cs.DS↗

arXiv2014

Hypergraph Markov Operators, Eigenvalues and Approximation Algorithms

The celebrated Cheeger's Inequality \cite{am85,a86} establishes a bound on the expansion of a graph via its spectrum. This inequality is central to a rich spectral theory of graphs, based on studying the eigenvalues and eigenvectors of the adjacency matrix (and other related matrices) of graphs. It has remained open to define a suitable spectral model for hypergraphs whose spectra can be used to estimate various combinatorial properties of the hypergraph. In this paper we introduce a new hypergraph Laplacian operator (generalizing the Laplacian matrix of graphs)and study its spectra. We prove a Cheeger-type inequality for hypergraphs, relating the second smallest eigenvalue of this operator to the expansion of the hypergraph. We bound other hypergraph expansion parameters via higher eigenvalues of this operator. We give bounds on the diameter of the hypergraph as a function of the second smallest eigenvalue of the Laplacian operator. The Markov process underlying the Laplacian operator can be viewed as a dispersion process on the vertices of the hypergraph that might be of independent interest. We bound the {\em Mixing-time} of this process as a function of the second smallest eigenvalue of the Laplacian operator. All these results are generalizations of the corresponding results for graphs. We show that there can be no linear operator for hypergraphs whose spectra captures hypergraph expansion in a Cheeger-like manner. For any $k$, we give a polynomial time algorithm to compute an approximation to the $k^{th}$ smallest eigenvalue of the operator. We show that this approximation factor is optimal under the SSE hypothesis (introduced by \cite{rs10}) for constant values of $k$. Finally, using the factor preserving reduction from vertex expansion in graphs to hypergraph expansion, we show that all our results for hypergraphs extend to vertex expansion in graphs.

cs.DM↗

arXiv2014

Approximation Algorithms for Hypergraph Small Set Expansion and Small Set Vertex Expansion

The expansion of a hypergraph, a natural extension of the notion of expansion in graphs, is defined as the minimum over all cuts in the hypergraph of the ratio of the number of the hyperedges cut to the size of the smaller side of the cut. We study the Hypergraph Small Set Expansion problem, which, for a parameter $δ\in (0,1/2]$, asks to compute the cut having the least expansion while having at most $δ$ fraction of the vertices on the smaller side of the cut. We present two algorithms. Our first algorithm gives an $\tilde O(δ^{-1} \sqrt{\log n})$ approximation. The second algorithm finds a set with expansion $\tilde O(δ^{-1}(\sqrt{d_{\text{max}}r^{-1}\log r\, ϕ^*} + ϕ^*))$ in a $r$--uniform hypergraph with maximum degree $d_{\text{max}}$ (where $ϕ^*$ is the expansion of the optimal solution). Using these results, we also obtain algorithms for the Small Set Vertex Expansion problem: we get an $\tilde O(δ^{-1} \sqrt{\log n})$ approximation algorithm and an algorithm that finds a set with vertex expansion $O\left(δ^{-1}\sqrt{ϕ^V \log d_{\text{max}} } + δ^{-1} ϕ^V\right)$ (where $ϕ^V$ is the vertex expansion of the optimal solution). For $δ=1/2$, Hypergraph Small Set Expansion is equivalent to the hypergraph expansion problem. In this case, our approximation factor of $O(\sqrt{\log n})$ for expansion in hypergraphs matches the corresponding approximation factor for expansion in graphs due to ARV.

cs.DS↗

arXiv2013

The Complexity of Approximating Vertex Expansion

We study the complexity of approximating the vertex expansion of graphs $G = (V,E)$, defined as \[ Φ^V := \min_{S \subset V} n \cdot \frac{|N(S)|}{|S| |V \backslash S|}. \] We give a simple polynomial-time algorithm for finding a subset with vertex expansion $O(\sqrt{OPT \log d})$ where $d$ is the maximum degree of the graph. Our main result is an asymptotically matching lower bound: under the Small Set Expansion (SSE) hypothesis, it is hard to find a subset with expansion less than $C\sqrt{OPT \log d}$ for an absolute constant $C$. In particular, this implies for all constant $ε> 0$, it is SSE-hard to distinguish whether the vertex expansion $< ε$ or at least an absolute constant. The analogous threshold for edge expansion is $\sqrt{OPT}$ with no dependence on the degree; thus our results suggest that vertex expansion is harder to approximate than edge expansion. In particular, while Cheeger's algorithm can certify constant edge expansion, it is SSE-hard to certify constant vertex expansion in graphs. Our proof is via a reduction from the {\it Unique Games} instance obtained from the \SSE hypothesis to the vertex expansion problem. It involves the definition of a smoother intermediate problem we call {\sf Analytic Vertex Expansion} which is representative of both the vertex expansion and the conductance of the graph. Both reductions (from the UGC instance to this problem and from this problem to vertex expansion) use novel proof ideas.

cs.CC↗

arXiv2013

Approximation Algorithm for Sparsest k-Partitioning

Given a graph $G$, the sparsest-cut problem asks to find the set of vertices $S$ which has the least expansion defined as $$ϕ_G(S) := \frac{w(E(S,\bar{S}))}{\min \set{w(S), w(\bar{S})}}, $$ where $w$ is the total edge weight of a subset. Here we study the natural generalization of this problem: given an integer $k$, compute a $k$-partition $\set{P_1, \ldots, P_k}$ of the vertex set so as to minimize $$ ϕ_k(\set{P_1, \ldots, P_k}) := \max_i ϕ_G(P_i). $$ Our main result is a polynomial time bi-criteria approximation algorithm which outputs a $(1 - \e)k$-partition of the vertex set such that each piece has expansion at most $O_{\varepsilon}(\sqrt{\log n \log k})$ times $OPT$. We also study balanced versions of this problem.

cs.DS↗

arXiv2011

Many Sparse Cuts via Higher Eigenvalues

Cheeger's fundamental inequality states that any edge-weighted graph has a vertex subset $S$ such that its expansion (a.k.a. conductance) is bounded as follows: \[ ϕ(S) \defeq \frac{w(S,\bar{S})}{\min \set{w(S), w(\bar{S})}} \leq 2\sqrt{λ_2} \] where $w$ is the total edge weight of a subset or a cut and $λ_2$ is the second smallest eigenvalue of the normalized Laplacian of the graph. Here we prove the following natural generalization: for any integer $k \in [n]$, there exist $ck$ disjoint subsets $S_1, ..., S_{ck}$, such that \[ \max_i ϕ(S_i) \leq C \sqrt{λ_{k} \log k} \] where $λ_i$ is the $i^{th}$ smallest eigenvalue of the normalized Laplacian and $c<1,C>0$ are suitable absolute constants. Our proof is via a polynomial-time algorithm to find such subsets, consisting of a spectral projection and a randomized rounding. As a consequence, we get the same upper bound for the small set expansion problem, namely for any $k$, there is a subset $S$ whose weight is at most a $\bigO(1/k)$ fraction of the total weight and $ϕ(S) \le C \sqrt{λ_k \log k}$. Both results are the best possible up to constant factors. The underlying algorithmic problem, namely finding $k$ subsets such that the maximum expansion is minimized, besides extending sparse cuts to more than one subset, appears to be a natural clustering problem in its own right.

cs.DS↗

arXiv2010

Cut-Matching Games on Directed Graphs

We give O(log^2 n)-approximation algorithm based on the cut-matching framework of [10, 13, 14] for computing the sparsest cut on directed graphs. Our algorithm uses only O(log^2 n) single commodity max-flow computations and thus breaks the multicommodity-flow barrier for computing the sparsest cut on directed graphs

cs.DS↗