Corpus ID: 237210690

The Weisfeiler-Leman Algorithm: An Exploration of its Power

@inproceedings{Kiefer2021TheWA,
  title={The Weisfeiler-Leman Algorithm: An Exploration of its Power},
  author={Sandra Kiefer},
  year={2021}
}
Some of my favorite open problems concern fixed-point logic with counting, FPC. It is known that counting logic with k + 1 variables, Ck+1, has exactly the same expressive power as the classic k-dimensional Weisfeiler-Leman Algorithm, k-WL. Furthermore, the quantifier-depth of a Ck+1 formula needed to express the color of a k-tuple of vertices is equal to the number of iterations of k-WL needed to derive that color. Much has been learned about FPC and the power of k-WL in the last forty years… Expand

Figures from this paper

Logarithmic Weisfeiler-Leman Identifies All Planar Graphs
TLDR
It is shown that there is a constant k such that every planar graph can be identified by the k-dimensional WL algorithm within a logarithmic number of iterations, which implies that everyPlanar graph is definable with a Ck+1-sentence of logarithsmic quantifier depth. Expand
The Power of the Weisfeiler-Leman Algorithm to Decompose Graphs
TLDR
It is proved that the 2-dimensional Weisfeiler-Leman algorithm implicitly computes the decomposition of a graph into its 3-connected components and insights are obtained about the connectivity of constituent graphs of association schemes. Expand
Lovász-Type Theorems and Game Comonads
TLDR
This work proposes a new categorical formulation, which applies to any locally finite category with pushouts and a proper factorisation system, and presents a novel application to homomorphism counts in modal logic. Expand
The Logic of Graph Neural Networks
  • Martin Grohe
  • Computer Science
  • 2021 36th Annual ACM/IEEE Symposium on Logic in Computer Science (LICS)
  • 2021
TLDR
This paper explains how the expressiveness of GNNs can be characterised precisely by the combinatorial Weisfeiler-Leman algorithms and by finite variable counting logics. Expand
Graph Learning with 1D Convolutions on Random Walks
TLDR
It is demonstrated empirically that CRAWL matches or outperforms state-of-the-art GNN architectures across a multitude of benchmark datasets for classification and regression on graphs. Expand
Recent Advances on the Graph Isomorphism Problem
TLDR
The main focus will be on Babai's quasi-polynomial time isomorphism test and subsequent developments that led to the design of isomorphicism algorithms with a quasi- polynomial parameterized running time of the from $n^{\polylog k}$, where $k$ is a graph parameter such as the maximum degree. Expand
Seurat games on Stockmeyer graphs
We define a family of vertex colouring games played over a pair of graphs or digraphs $(G,H)$ by players $\forall$ and $\exists$. These games arise from work on a longstanding open problem inExpand

References

SHOWING 1-10 OF 41 REFERENCES
A Linear Upper Bound on the Weisfeiler-Leman Dimension of Graphs of Bounded Genus
TLDR
It is proved that the WL dimension of graphs embeddable in a surface of Euler genus g is at most $4g+3$ and the approach yields an upper bound of $2g-3. Expand
Weisfeiler-Lehman Refinement Requires at Least a Linear Number of Iterations
TLDR
It is shown that for constants h ≥ 1, there are pairs of graphs such that h-dimensional Weisfeiler-Lehman refinement (h-dim W-L) can distinguish the two graphs, but requires at least a linear number of iterations. Expand
Walk refinement, walk logic, and the iteration number of the Weisfeiler-Leman algorithm
TLDR
This work exploits a new refinement based on counting walks and argues that its iteration number differs from the classic Weisfeiler-Leman refinement by at most a logarithmic factor, and proves matching linear upper and lower bounds on the number of iterations of the walk refinement. Expand
Identifiability of Graphs with Small Color Classes by the Weisfeiler-Leman Algorithm
TLDR
An efficient decision procedure is given that recognizes whether or not a graph $G$ of color multiplicity 4 is identifiable by 2-WL, that is,Whether or not 2- WL distinguishes $G $ from any non-isomorphic graph. Expand
Upper Bounds on the Quantifier Depth for Graph Differentiation in First Order Logic
TLDR
It is shown that on graphs with n vertices the 2-dimensional Weisfei-ler-Leman algorithm requires at most O(n2/log(n) iterations to reach stabilization, and the previously best, trivial upper bound of O( n2) is asymptotically not tight. Expand
The Power of the Weisfeiler-Leman Algorithm to Decompose Graphs
TLDR
It is proved that the 2-dimensional Weisfeiler-Leman algorithm implicitly computes the decomposition of a graph into its 3-connected components and insights are obtained about the connectivity of constituent graphs of association schemes. Expand
Sherali-Adams relaxations and indistinguishability in counting logics
TLDR
It is shown that certain properties of graphs, such as that of having a flow-circulation of a prescribed value, are definable in the infinitary logic with counting with a bounded number of variables. Expand
Universal Covers, Color Refinement, and Two-Variable Counting Logic: Lower Bounds for the Depth
TLDR
This work establishes that color refinement, the classical procedure used in isomorphism testing and other areas for computing the coarsest equitable partition of a graph, needs (1-o(1))n rounds to achieve color stabilization on each of G and H. Expand
Isomorphism testing for embeddable graphs through definability
TLDR
It is shown that for every surface S, there is a k _> 1 such that the k-dimensional WL-algori thm succeeds to decide isomorphism of graphs embeddable in S. Expand
Pebble Games and Linear Equations
TLDR
A precise characterisation of the level-k Sherali-Adams relaxation in terms of a modified counting pebble game is given and a proof that the interleaving between these two hierarchies is strict is proved. Expand
...
1
2
3
4
5
...