Abstract:Machine learning is becoming an increasingly valuable tool in mathematics, enabling one to identify subtle patterns across collections of examples so vast that they would be impossible for a single researcher to feasibly review and analyze. In this work, we use graph neural networks to investigate quiver mutation -- an operation that transforms one quiver (or directed multigraph) into another -- which is central to the theory of cluster algebras with deep connections to geometry, topology, and physics. In the study of cluster algebras, the question of mutation equivalence is of fundamental concern: given two quivers, can one efficiently determine if one quiver can be transformed into the other through a sequence of mutations? Currently, this question has only been resolved in specific cases. In this paper, we use graph neural networks and AI explainability techniques to discover mutation equivalence criteria for the previously unknown case of quivers of type $\tilde{D}_n$. Along the way, we also show that even without explicit training to do so, our model captures structure within its hidden representation that allows us to reconstruct known criteria from type $D_n$, adding to the growing evidence that modern machine learning models are capable of learning abstract and general rules from mathematical data.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: **How to efficiently determine whether two quivers can be transformed into each other through a series of mutation operations, especially for the problem of mutation - equivalent classes of previously unknown $\tilde{D}_n$ - type quivers.** Specifically, the paper focuses on quiver mutation, an operation that transforms one quiver into another and plays a central role in the theory of cluster algebras and has profound connections with fields such as geometry, topology, and physics. In the study of cluster algebras, the mutation - equivalence problem is fundamental: given two quivers, can one effectively determine whether one quiver can be transformed into the other through a series of mutation operations. Currently, this problem has only been solved in specific cases. This paper uses graph neural networks (GNNs) and AI interpretability techniques to discover the mutation - equivalence criteria for $\tilde{D}_n$ - type quivers. In addition, the research also shows that even without explicitly training the model for this task, the structures captured by the model in its hidden representations can reconstruct the known mutation - equivalence criteria for $D_n$ - type quivers, further demonstrating the ability of modern machine - learning models to learn abstract and general rules from mathematical data. ### The main contributions of the paper include: 1. **Application of graph neural networks**: Used to represent the mutation - equivalent classes of quivers. 2. **Model interpretability**: Provides evidence that the model has learned features consistent with the known mutation - equivalence criteria for $D_n$ - type quivers. 3. **Proof of new theorems**: Through the interpretation of the model, a brand - new theorem is proved, which fully characterizes the mutation - equivalent classes of $\tilde{D}_n$ - type quivers. These contributions not only demonstrate the potential of machine learning in mathematical research but also provide new tools and methods for further understanding quiver mutation - equivalent classes.

Machines and Mathematical Mutations: Using GNNs to Characterize Quiver Mutation Classes

Machine Learning Mutation-Acyclicity of Quivers

Quiver mutations, Seiberg duality, and machine learning

Cluster Algebras: Network Science and Machine Learning

Mutation commutativity between geometric realizations and related quivers of classical cluster algebras and applications

The Representation Theory of Neural Networks

Theoretical Guarantees for Permutation-Equivariant Quantum Neural Networks

A topology on the poset of quiver mutation classes

Geometry of $C$-vectors and $C$-Matrices for Mutation-Infinite Quivers

MuNN: Mutation Analysis of Neural Networks

Mutation of frozen Jacobian algebras

Problem-informed Graphical Quantum Generative Learning

Applications of Mutations in the Derived Categories of Weighted Projective Lines to Lie and Quantum Algebras

Hidden Activations Are Not Enough: A General Approach to Neural Network Predictions

ALPHAGMUT: A Rationale-Guided Alpha Shape Graph Neural Network to Evaluate Mutation Effects

Recent Advances for Quantum Neural Networks in Generative Learning

Equivariant quantum circuits for learning on weighted graphs

Geometric quantum machine learning of BQP$^A$ protocols and latent graph classifiers

Geometrically Equivariant Graph Neural Networks: A Survey

The coloured mutation class of $\mathbb{A}_n$ quivers