Abstract:We develop further the graph limit theory for dense weighted graph sequences. In particular, we consider probability graphons, which have recently appeared in graph limit theory as continuum representations of weighted graphs, and we introduce P-variables, which also appear in the context of the Aldous-Hoover theorem for exchangeable infinite random arrays, as an alternative continuum representation for weighted graphs. In particular, we explain how P-variables are related to probability graphons in a similar way in which random variables are related to probability measures. We define a metric for P-variables (inspired by action convergence in the graph limit theory of sparse graph sequences) and show that convergence of P-variables in this metric is equivalent to probability graphons convergence. We exploit this equivalence to translate several results from the theory of probability graphons to P-variables. In addition, we prove several properties of P-variables convergence, thus showing new properties also for probability graphons convergence and demonstrating the power of the connection between probability graphons and P-variables. Furthermore, we show how P-variables convergence can be easily modified and generalised to cover other combinatorial structures such as bipartite graphs and hypergraphs.

What problem does this paper attempt to address?

The paper attempts to address the following problem: further developing the equivalence relationship between probability graphons and P-variables in the limit theory of dense weighted graph sequences. Specifically, the authors aim to introduce P-variables as an alternative continuous representation method for weighted graphs and demonstrate the relationship between P-variables and probability graphons, thereby providing new perspectives and tools for the limit theory of weighted graphs. ### Main Issues 1. **Relationship between Probability Graphons and P-variables**: - The paper explores the relationship between P-variables and probability graphons, similar to the relationship between random variables and probability measures. - The authors define a pseudometric for P-variables and prove that this metric is topologically equivalent to the unlabeled cut metric of probability graphons on compact subsets. 2. **Convergence**: - The authors prove that the convergence of P-variables is equivalent to the convergence of probability graphons. - Through this equivalence, the authors translate several results from probability graphons into the language of P-variables and prove new convergence properties for P-variables. 3. **Application Extensions**: - The authors demonstrate how the convergence of P-variables can be applied to other combinatorial structures, such as bipartite graphs and hypergraphs. - Using the framework of P-variables, the authors provide a systematic and transparent method to define the limits of hypergraphs and explain the connections with existing work. ### Specific Goals - **Define P-variables**: Introduce P-variables as a continuous representation method for weighted graphs. - **Establish Equivalence**: Prove the equivalence between the convergence of P-variables and the convergence of probability graphons. - **Translate Existing Results**: Translate known results from probability graphons into the language of P-variables and prove new convergence properties for P-variables. - **Extend Applications**: Demonstrate the application of P-variables in bipartite graphs, self-loop graphs, sparse graph sequences, and hypergraph sequences. ### Methods - **Define P-variables**: Define P-variables as measurable functions from a probability space to the set of real numbers. - **Define Pseudometric**: Define a pseudometric for P-variables and prove its equivalence to the unlabeled cut metric of probability graphons. - **Prove Convergence**: Prove the equivalence between the convergence of P-variables and the convergence of probability graphons through the established equivalence relationship. - **Application Extensions**: Demonstrate the application of P-variables in different combinatorial structures and provide specific examples. ### Significance - **Theoretical Contribution**: Provide new perspectives and tools for the limit theory of weighted graphs, enriching the study of graph limit theory. - **Practical Application**: The framework of P-variables can be applied to complex network analysis, graph neural networks, and other fields, providing a stronger mathematical foundation. Through this research, the paper provides a new theoretical foundation and practical tools for the limit theory of weighted graphs, promoting further development in related fields.

Probability graphons and P-variables: two equivalent viewpoints for dense weighted graph limits

Probability graphons: the right convergence point of view

On the convergence of probabilities of the random graphs' properties expressed by first-order formulae with a bounded quantifier depth

An $L^p$ theory of sparse graph convergence I: Limits, sparse random graph models, and power law distributions

Random minimum spanning tree and dense graph limits

Graphon-Theoretic Approach to Central Limit Theorems for $ε$-Independence

On exchangeable random variables and the statistics of large graphs and hypergraphs

General limit theorems for mixtures of free, monotone, and boolean independence

Limiting Spectra of inhomogeneous random graphs

Ordered Graph Limits and Their Applications

Limiting Behavior of Weighted Sums of Negatively Associated Random Variables

Dense and nondense limits for uniform random intersection graphs

Local limit theorem for joint subgraph counts

Consistency of empirical distributions of sequences of graph statistics in networks with dependent edges

Convergent sequences of sparse graphs: A large deviations approach

Action convergence of general hypergraphs and tensors

Last passage percolation and limit theorems in Barak-Erdős directed random graphs and related models

Undecidability of polynomial inequalities in weighted graph homomorphism densities

Graphon Particle Systems, Part I: Spatio-Temporal Approximation and Law of Large Numbers

Coarse embeddability, $L^1$-compression and Percolations on General Graphs

Path convergence of Markov chains on large graphs