Erasure-Resilient Sublinear-Time Graph Algorithms

We investigate sublinear-time algorithms that take partially erased graphs represented by adjacency lists as input. Our algorithms make degree and neighbor queries to the input graph and work with a specified fraction of adversarial erasures in adjacency entries. We focus on two computational tasks: testing if a graph is connected or ε-far from connected and estimating the average degree. For testing connectedness, we discover a threshold phenomenon: when the fraction of erasures is less than ε, this property can be tested efficiently (in time independent of the size of the graph); when the fraction of erasures is at least ε, then a number of queries linear in the size of the graph representation is required. Our erasure-resilient algorithm (for the special case with no erasures) is an improvement over the previously known algorithm for connectedness in the standard property testing model and has optimal dependence on the proximity parameter ε. For estimating the average degree, our results provide an “interpolation” between the query complexity for this computational task in the model with no erasures in two different settings: with only degree queries, investigated by Feige (SIAM J. Comput. ‘06), and with degree queries and neighbor queries, investigated by Goldreich and Ron (Random Struct. Algorithms ‘08) and Eden et al. (ICALP ‘17). We conclude with a discussion of our model and open questions raised by our work.

Computing Graph Neural Networks: A Survey from Algorithms to Accelerators

Graph Neural Networks (GNNs) have exploded onto the machine learning scene in recent years owing to their capability to model and learn from graph-structured data. Such an ability has strong implications in a wide variety of fields whose data are inherently relational, for which conventional neural networks do not perform well. Indeed, as recent reviews can attest, research in the area of GNNs has grown rapidly and has lead to the development of a variety of GNN algorithm variants as well as to the exploration of ground-breaking applications in chemistry, neurology, electronics, or communication networks, among others. At the current stage research, however, the efficient processing of GNNs is still an open challenge for several reasons. Besides of their novelty, GNNs are hard to compute due to their dependence on the input graph, their combination of dense and very sparse operations, or the need to scale to huge graphs in some applications. In this context, this article aims to make two main contributions. On the one hand, a review of the field of GNNs is presented from the perspective of computing. This includes a brief tutorial on the GNN fundamentals, an overview of the evolution of the field in the last decade, and a summary of operations carried out in the multiple phases of different GNN algorithm variants. On the other hand, an in-depth analysis of current software and hardware acceleration schemes is provided, from which a hardware-software, graph-aware, and communication-centric vision for GNN accelerators is distilled.

Counting Induced Subgraphs: An Algebraic Approach to #W[1]-Hardness

We study the problem $$\#\textsc {IndSub}(\varPhi )$$ # I N D S U B ( Φ ) of counting all induced subgraphs of size k in a graph G that satisfy the property $$\varPhi $$ Φ . It is shown that, given any graph property $$\varPhi $$ Φ that distinguishes independent sets from bicliques, $$\#\textsc {IndSub}(\varPhi )$$ # I N D S U B ( Φ ) is hard for the class $$\#\mathsf {W[1]}$$ # W [ 1 ] , i.e., the parameterized counting equivalent of $${{\mathsf {N}}}{{\mathsf {P}}}$$ N P . Under additional suitable density conditions on $$\varPhi $$ Φ , satisfied e.g. by non-trivial monotone properties on bipartite graphs, we strengthen $$\#\mathsf {W[1]}$$ # W [ 1 ] -hardness by establishing that $$\#\textsc {IndSub}(\varPhi )$$ # I N D S U B ( Φ ) cannot be solved in time $$f(k)\cdot n^{o(k)}$$ f ( k ) · n o ( k ) for any computable function f, unless the Exponential Time Hypothesis fails. Finally, we observe that our results remain true even if the input graph G is restricted to be bipartite and counting is done modulo a fixed prime.

Molecular Graph Contrastive Learning with Parameterized Explainable Augmentations

ABSTRACTLearning generalizable, transferable, and robust representations for molecule data has always been a challenge. The recent success of contrastive learning (CL) for self-supervised graph representation learning provides a novel perspective to learn molecule representations. The most prevailing graph CL framework is to maximize the agreement of representations in different augmented graph views. However, existing graph CL frameworks usually adopt stochastic augmentations or schemes according to pre-defined rules on the input graph to obtain different graph views in various scales (e.g. node, edge, and subgraph), which may destroy topological semantemes and domain prior in molecule data, leading to suboptimal performance. Therefore, designing parameterized, learnable, and explainable augmentation is quite necessary for molecular graph contrastive learning. A well-designed parameterized augmentation scheme can preserve chemically meaningful structural information and intrinsically essential attributes for molecule graphs, which helps to learn representations that are insensitive to perturbation on unimportant atoms and bonds. In this paper, we propose a novel Molecular Graph Contrastive Learning with Parameterized Explainable Augmentations, MolCLE for brevity, that self-adaptively incorporates chemically significative information from both topological and semantic aspects of molecular graphs. Specifically, we apply deep neural networks to parameterize the augmentation process for both the molecular graph topology and atom attributes, to highlight contributive molecular substructures and recognize underlying chemical semantemes. Comprehensive experiments on a variety of real-world datasets demonstrate that our proposed method consistently outperforms compared baselines, which verifies the effectiveness of the proposed framework. Detailedly, our self-supervised MolCLE model surpasses many supervised counterparts, and meanwhile only uses hundreds of thousands of parameters to achieve comparative results against the state-of-the-art baseline, which has tens of millions of parameters. We also provide detailed case studies to validate the explainability of augmented graph views.CCS CONCEPTS• Mathematics of computing → Graph algorithms; • Applied computing → Bioinformatics; • Computing methodologies → Neural networks; Unsupervised learning.

Structural Parameterizations of Clique Coloring

A clique coloring of a graph is an assignment of colors to its vertices such that no maximal clique is monochromatic. We initiate the study of structural parameterizations of the Clique Coloring problem which asks whether a given graph has a clique coloring with q colors. For fixed $$q \ge 2$$ q ≥ 2 , we give an $$\mathscr {O}^{\star }(q^{{\mathsf {tw}}})$$ O ⋆ ( q tw ) -time algorithm when the input graph is given together with one of its tree decompositions of width $${\mathsf {tw}} $$ tw . We complement this result with a matching lower bound under the Strong Exponential Time Hypothesis. We furthermore show that (when the number of colors is unbounded) Clique Coloring is $$\mathsf {XP}$$ XP parameterized by clique-width.

Better Distance Preservers and Additive Spanners

We study two popular ways to sketch the shortest path distances of an input graph. The first is distance preservers , which are sparse subgraphs that agree with the distances of the original graph on a given set of demand pairs. Prior work on distance preservers has exploited only a simple structural property of shortest paths, called consistency , stating that one can break shortest path ties such that no two paths intersect, split apart, and then intersect again later. We prove that consistency alone is not enough to understand distance preservers, by showing both a lower bound on the power of consistency and a new general upper bound that polynomially surpasses it. Specifically, our new upper bound is that any p demand pairs in an n -node undirected unweighted graph have a distance preserver on O( n 2/3 p 2/3 + np 1/3 edges. We leave a conjecture that the right bound is O ( n 2/3 p 2/3 + n ) or better. The second part of this paper leverages these distance preservers in a new construction of additive spanners , which are subgraphs that preserve all pairwise distances up to an additive error function. We give improved error bounds for spanners with relatively few edges; for example, we prove that all graphs have spanners on O(n) edges with + O ( n 3/7 + ε ) error. Our construction can be viewed as an extension of the popular path-buying framework to clusters of larger radii.

Multiplicative Parameterization Above a Guarantee

Parameterization above a guarantee is a successful paradigm in Parameterized Complexity. To the best of our knowledge, all fixed-parameter tractable problems in this paradigm share an additive form defined as follows. Given an instance ( I,k ) of some (parameterized) problem π with a guarantee g(I) , decide whether I admits a solution of size at least (or at most) k + g(I) . Here, g(I) is usually a lower bound on the minimum size of a solution. Since its introduction in 1999 for M AX SAT and M AX C UT (with g(I) being half the number of clauses and half the number of edges, respectively, in the input), analysis of parameterization above a guarantee has become a very active and fruitful topic of research. We highlight a multiplicative form of parameterization above (or, rather, times) a guarantee: Given an instance ( I,k ) of some (parameterized) problem π with a guarantee g(I) , decide whether I admits a solution of size at least (or at most) k · g(I) . In particular, we study the Long Cycle problem with a multiplicative parameterization above the girth g(I) of the input graph, which is the most natural guarantee for this problem, and provide a fixed-parameter algorithm. Apart from being of independent interest, this exemplifies how parameterization above a multiplicative guarantee can arise naturally. We also show that, for any fixed constant ε > 0, multiplicative parameterization above g(I) 1+ε of Long Cycle yields para-NP-hardness, thus our parameterization is tight in this sense. We complement our main result with the design (or refutation of the existence) of fixed-parameter algorithms as well as kernelization algorithms for additional problems parameterized multiplicatively above girth.

Reasoning about Explanations for Non-validation in SHACL

The Shapes Constraint Language (SHACL) is a recently standardized language for describing and validating constraints over RDF graphs. The SHACL specification describes the so-called validation reports, which are meant to explain to the users the outcome of validating an RDF graph against a collection of constraints. Specifically, explaining the reasons why the input graph does not satisfy the constraints is challenging. In fact, the current SHACL standard leaves it open on how such explanations can be provided to the users. In this paper, inspired by works on logic-based abduction and database repairs, we study the problem of explaining non-validation of SHACL constraints. In particular, in our framework non-validation is explained using the notion of a repair, i.e., a collection of additions and deletions whose application on an input graph results in a repaired graph that does satisfy the given SHACL constraints. We define a collection of decision problems for reasoning about explanations, possibly restricting to explanations that are minimal with respect to cardinality or set inclusion. We provide a detailed characterization of the computational complexity of those reasoning tasks, including the combined and the data complexity.

Multi-Scale Contrastive Siamese Networks for Self-Supervised Graph Representation Learning

Graph representation learning plays a vital role in processing graph-structured data. However, prior arts on graph representation learning heavily rely on labeling information. To overcome this problem, inspired by the recent success of graph contrastive learning and Siamese networks in visual representation learning, we propose a novel self-supervised approach in this paper to learn node representations by enhancing Siamese self-distillation with multi-scale contrastive learning. Specifically, we first generate two augmented views from the input graph based on local and global perspectives. Then, we employ two objectives called cross-view and cross-network contrastiveness to maximize the agreement between node representations across different views and networks. To demonstrate the effectiveness of our approach, we perform empirical experiments on five real-world datasets. Our method not only achieves new state-of-the-art results but also surpasses some semi-supervised counterparts by large margins. Code is made available at https://github.com/GRAND-Lab/MERIT

On star family packing of graphs

Let $\mathcal{H}$ be a family of graphs. An $\mathcal{H}$-packing of a graph $G$ is a set $\{G_1,G_2,\dots,G_k\}$ of disjoint subgraphs of $G$ such that each $G_j$ is isomorphic to some element of $\mathcal{H}$. An $\mathcal{H}$-packing of a graph $G$ that covers the maximum number of vertices of $G$ is called a maximum $\mathcal{H}$-packing of $G$. The $\mathcal{H}$-packing problem seeks to find a maximum $\mathcal{H}$-packing of a graph. Let $i$ be a positive integer. An $i$-star is a complete bipartite graph $K_{1,i}$. This paper investigates the $\mathcal{H}$-packing problem with $\mathcal{H}$ being a family of stars. For an arbitrary family $\mathcal{S}$ of stars, we design a linear-time algorithm for the $\mathcal{S}$-packing problem in trees. Let $t$ be a positive integer. An $\mathcal{H}$-packing is called a $t^+$-star packing if $\mathcal{H}$ consists of all $i$-stars with $i\ge t$. We show that the $t^+$-star packing problem for $t\ge 2$ is NP-hard in bipartite graphs. As a consequence, the $2^+$-star packing problem is NP-hard even in bipartite graphs with maximum degree at most $4$. Let $T$ and $t$ be two positive integers with $T>t$. An $\mathcal{H}$-packing is called a $T\setminus t$-star packing if $\mathcal{H}=\{K_{1,1},K_{1,2},\dots,K_{1,T}\}\setminus \{K_{1,t}\}$. For $t\ge 2$, we present a $\frac{t}{t+1}$-approximation algorithm for the $T\setminus t$-star packing problem that runs in $\mathcal{O}(mn^{1/2})$ time, where $n$ is the number of vertices and $m$ the number of edges of the input graph. We also design a $\frac{1}{2}$-approximation algorithm for the $2^+$-star packing problem that runs in $\mathcal{O}(m)$ time, where $m$ is the number of edges of the input graph. As a consequence, every connected graph with at least $3$ vertices has a $2^+$-star packing that covers at least half of its vertices.