Error-Correcting Graph Codes

Swastik Kopparty,Aditya Potukuchi,Harry Sha

2024-10-09

Abstract:In this paper, we construct Error-Correcting Graph Codes. An error-correcting graph code of distance $\delta$ is a family $C$ of graphs on a common vertex set of size $n$, such that if we start with any graph in $C$, we would have to modify the neighborhoods of at least $\delta n$ vertices in order to obtain some other graph in $C$. This is a natural graph generalization of the standard Hamming distance error-correcting codes for binary strings. Yohananov and Yaakobi were the first to construct codes in this metric, constructing good codes for $\delta < 1/2$, and optimal codes for a large-alphabet analogue. We extend their work by showing 1. Combinatorial results determining the optimal rate vs. distance trade-off nonconstructively. 2. Graph code analogues of Reed-Solomon codes and code concatenation, leading to positive distance codes for all rates and positive rate codes for all distances. 3. Graph code analogues of dual-BCH codes, yielding large codes with distance $\delta = 1-o(1)$. This gives an explicit ''graph code of Ramsey graphs''. Several recent works, starting with the paper of Alon, Gujgiczer, Körner, Milojević, and Simonyi, have studied more general graph codes; where the symmetric difference between any two graphs in the code is required to have some desired property. Error-correcting graph codes are a particularly interesting instantiation of this concept.

Information Theory,Discrete Mathematics,Combinatorics

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to construct Error - Correcting Graph Codes in graph theory. Specifically, the authors hope to construct a family of graph codes \(C\), which are defined on the same vertex set, and the "distance" between any two graph codes is at least \(\delta n\). Here, "distance" means that at least \(\delta n\) vertex neighborhoods need to be modified to transform from one graph to another. This concept of distance is a natural generalization of the standard Hamming - distance error - correcting codes in graph theory.

### Main research objectives
1. **Determine the optimal rate - distance trade - off**: Determine the optimal rate - distance trade - off relationship of graph codes non - constructively through combinatorial methods.
2. **Construct graph codes with positive distances**: Provide construction methods for graph codes with positive distances for all rates and all distances.
3. **Construct graph codes with high distances**: In particular, for graph codes with distance \(\delta = 1 - o(1)\), provide explicit construction methods.

### Specific problems
- **How to construct graph codes with high error - correction capabilities**: In particular, in the binary case, construct graph codes with a distance greater than 1/2 and a constant rate.
- **How to optimize the construction methods of graph codes**: Provide explicit and strongly - explicit construction methods for graph codes to ensure efficiency and feasibility in practical applications.
- **Explore the relationship between graph codes and pseudo - random graphs**: Study the connection between graph codes and Ramsey graphs, and their applications in pseudo - randomness.

### Research background
- **Previous work**: Yohananov and Yaakobi first constructed graph codes under this metric and provided good construction methods when \(\delta < 1/2\). They also studied the optimal construction methods in the case of large alphabets.
- **Related research**: Alon et al. studied more general graph codes, in which the symmetric difference between any two graph codes needs to satisfy certain specific properties. Error - correcting graph codes are a particularly interesting instance of this concept.

### Main contributions
1. **Non - constructive existence results**: Prove the existence of graph codes with an optimal rate - distance trade - off relationship.
2. **Explicit construction methods**:
- Provide a construction method based on tensor products and zero - diagonal symmetric matrices (STCZD), which makes graph codes have good distance characteristics.
- Use Reed - Solomon codes and code - splicing techniques to construct graph codes with high distances and positive rates.
- For very high distances \(\delta = 1 - O(n^{-\epsilon})\), provide an explicit construction method to obtain "Ramsey graphs of graph codes".

### Technical means
- **Probability method**: Prove the existence of optimal graph codes through the method of random linear subspaces.
- **Tensor products and zero - diagonal symmetric matrices**: Use tensor products and zero - diagonal symmetric matrices to convert classical Hamming - distance codes into graph codes.
- **Code - splicing technique**: Through multi - layer code - splicing, gradually improve the distance and rate of graph codes.
- **Construction of Ramsey graphs**: Use polynomials over finite fields to construct graph codes with high distances, which are of great significance in the study of Ramsey graphs.

### Conclusion
This paper not only theoretically determines the optimal rate - distance trade - off relationship of graph codes, but also provides a variety of explicit construction methods. In particular, in the binary case, it constructs graph codes with a distance greater than 1/2 and a constant rate. These results provide a theoretical basis and technical support for the efficient implementation of graph codes in practical applications.