F. Constantino

Y. He

E. Heyes

E. Hirst

3-manifolds represented as isomorphism signatures of their triangulations and associated Pachner graphs are analysed with machine learning. 

We apply machine learning to the study of topological features related to Sasakian and G2-geometries derived from Calabi-Yau manifolds. We compute extensive datasets of invariants—including Hodge numbers and the Crowley–Nördstrom invariant—for more than 7000 manifolds realised as hypersurfaces in weighted projective space. Neural networks and a symbolic regressors are used to uncover new patterns in these invariants.

Real 3-manifold triangulations can be uniquely represented by isomorphism signatures. Databases of these isomorphism signatures are generated for a variety of 3-manifolds and knot complements, using SnapPy and Regina, then these language-like inputs are used to train various machine learning architectures to differentiate the manifolds, as well as their Dehn surgeries, via their triangulations. Gradient saliency analysis then extracts key parts of this language-like encoding scheme from the trained models. The isomorphism signature databases are taken from the 3-manifolds' Pachner graphs, which are also generated in bulk for some selected manifolds of focus and for the subset of the SnapPy orientable cusped census with 
 initial tetrahedra. These Pachner graphs are further analysed through the lens of network science to identify new structure in the triangulation representation; in particular for the hyperbolic case, a relation between the length of the shortest geodesic (systole) and the size of the Pachner graph's ball is observed.

Learning 3-manifolds

Learning 3-manifold triangulations