Hypergraph Isomorphism Computation

Yifan Feng,Jiashu Han,Shihui Ying,Yue Gao

2023-07-26

Abstract:The isomorphism problem is a fundamental problem in network analysis, which involves capturing both low-order and high-order structural information. In terms of extracting low-order structural information, graph isomorphism algorithms analyze the structural equivalence to reduce the solver space dimension, which demonstrates its power in many applications, such as protein design, chemical pathways, and community detection. For the more commonly occurring high-order relationships in real-life scenarios, the problem of hypergraph isomorphism, which effectively captures these high-order structural relationships, cannot be straightforwardly addressed using graph isomorphism methods. Besides, the existing hypergraph kernel methods may suffer from high memory consumption or inaccurate sub-structure identification, thus yielding sub-optimal performance. In this paper, to address the abovementioned problems, we first propose the hypergraph Weisfiler-Lehman test algorithm for the hypergraph isomorphism test problem by generalizing the Weisfiler-Lehman test algorithm from graphs to hypergraphs. Secondly, based on the presented algorithm, we propose a general hypergraph Weisfieler-Lehman kernel framework and implement two instances, which are Hypergraph Weisfeiler-Lehamn Subtree Kernel and Hypergraph Weisfeiler-Lehamn Hyperedge Kernel. In order to fulfill our research objectives, a comprehensive set of experiments was meticulously designed, including seven graph classification datasets and 12 hypergraph classification datasets. Results on hypergraph classification datasets show significant improvements compared to other typical kernel-based methods, which demonstrates the effectiveness of the proposed methods. In our evaluation, we found that our proposed methods outperform the second-best method in terms of runtime, running over 80 times faster when handling complex hypergraph structures.

Data Structures and Algorithms,Artificial Intelligence,Machine Learning

What problem does this paper attempt to address?

### Problems Addressed by the Paper

The paper primarily aims to address the problem of isomorphism testing for higher-order structures (such as hypergraphs) and proposes a new hypergraph isomorphism testing algorithm along with a corresponding kernel function method.

#### Main Issues
1. **Limitations of Traditional Graph Isomorphism Algorithms**:
- Traditional graph isomorphism algorithms (such as the Weisfeiler-Lehman test) can only handle low-order structural information and cannot effectively capture higher-order relationships in the real world.

2. **Shortcomings of Existing Hypergraph Kernel Methods**:
- Existing hypergraph kernel methods may face issues such as high memory consumption or inaccurate substructure recognition, leading to poor performance.

#### Solutions
1. **Extending the Weisfeiler-Lehman Test to Hypergraphs**:
- A hypergraph Weisfeiler-Lehman test algorithm is proposed, which extends the Weisfeiler-Lehman test from graphs to hypergraphs to address the isomorphism testing problem for higher-order structures.

2. **Proposing a New Hypergraph Kernel Framework**:
- Based on the above algorithm, a general hypergraph Weisfeiler-Lehman kernel framework is proposed, and two instances are implemented: the hypergraph Weisfeiler-Lehman subtree kernel and the hypergraph Weisfeiler-Lehman hyperedge kernel.

3. **Mathematical Proof and Experimental Validation**:
- It is mathematically proven that the proposed hypergraph Weisfeiler-Lehman subtree kernel can degenerate into the classical graph Weisfeiler-Lehman subtree kernel when dealing with low-order graph structures.
- The effectiveness of the proposed methods is validated through extensive experiments, including 7 graph classification datasets and 12 hypergraph classification datasets.

#### Main Contributions
1. Extended the Weisfeiler-Lehman test algorithm to hypergraphs and proposed the hypergraph Weisfeiler-Lehman test algorithm.
2. Proposed a general hypergraph Weisfeiler-Lehman kernel framework and implemented two instances.
3. Mathematically proved the equivalence of the hypergraph Weisfeiler-Lehman subtree kernel when handling graph structures.
4. Extensive experiments demonstrated the effectiveness and speed advantages of the proposed methods.