REVIEW 10 cited by
Topological Deep Learning: Going Beyond Graph Data
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Topological deep learning is a rapidly growing field that pertains to the development of deep learning models for data supported on topological domains such as simplicial complexes, cell complexes, and hypergraphs, which generalize many domains encountered in scientific computations. In this paper, we present a unifying deep learning framework built upon a richer data structure that includes widely adopted topological domains. Specifically, we first introduce combinatorial complexes, a novel type of topological domain. Combinatorial complexes can be seen as generalizations of graphs that maintain certain desirable properties. Similar to hypergraphs, combinatorial complexes impose no constraints on the set of relations. In addition, combinatorial complexes permit the construction of hierarchical higher-order relations, analogous to those found in simplicial and cell complexes. Thus, combinatorial complexes generalize and combine useful traits of both hypergraphs and cell complexes, which have emerged as two promising abstractions that facilitate the generalization of graph neural networks to topological spaces. Second, building upon combinatorial complexes and their rich combinatorial and algebraic structure, we develop a general class of message-passing combinatorial complex neural networks (CCNNs), focusing primarily on attention-based CCNNs. We characterize permutation and orientation equivariances of CCNNs, and discuss pooling and unpooling operations within CCNNs in detail. Third, we evaluate the performance of CCNNs on tasks related to mesh shape analysis and graph learning. Our experiments demonstrate that CCNNs have competitive performance as compared to state-of-the-art deep learning models specifically tailored to the same tasks. Our findings demonstrate the advantages of incorporating higher-order relations into deep learning models in different applications.
Forward citations
Cited by 10 Pith papers
-
Foundations of Equivariant Deep Learning: Unifying Graph and Sheaf Neural Networks
Order-equivariant networks on poset bundles unify graph and sheaf message passing and are dense in continuous order-equivariant maps, with a further categorical generalization.
-
Differentiable Lifting for Topological Neural Networks
A differentiable lifting framework that samples and accepts candidate higher-order cells end-to-end outperforms static liftings on multiple TNN benchmarks.
-
Cosmology with Topological Deep Learning
Topological neural networks using tetrahedra, clusters and hyperedges built from halo catalogs lower inference error on Omega_m by 22% and on sigma_8 by up to 60% versus graph neural networks on Quijote.
-
Enhancing the Utility of Higher-Order Information in Relational Learning
Graph-level GNNs with new hypergraph-based encodings beat hypergraph-specific GNNs on several benchmarks, and the encodings provably increase expressivity beyond graph-level encodings.
-
New class of exactly flat topological bands - compact localised states protected by local graph topology
Face-vertex incidence matrices of arbitrary graphs generate exactly flat bands with degeneracy at least the number of faces minus the number of vertices.
-
Heat Kernel Goes Topological
TopoHKS defines a weighted combinatorial-complex Laplacian and heat kernel descriptor, claiming maximal expressive power; the supporting uniqueness theorem is incorrect.
-
Demystifying Topological Message-Passing with Relational Structures: A Case Study on Oversquashing in Simplicial Message-Passing
Simplicial message passing can be analyzed for oversquashing by collapsing its relational structure into an influence graph and applying graph-theoretic sensitivity, curvature, and rewiring tools.
-
CellCLAT: Preserving Topology and Trimming Redundancy in Self-Supervised Cellular Contrastive Learning
CellCLAT applies parameter-perturbation contrastive learning to cellular complexes and adaptively trims 2-cells to improve downstream graph classification.
-
Don't be Afraid of Cell Complexes! An Introduction from an Applied Perspective
A tutorial that defines cell complexes through boundary matrices and simple cycles, proves this matches topological regular cell complexes up to dimension two, and surveys signal processing and learning applications.
-
A Sheaf-Theoretic and Topological Perspective on Complex Network Modeling and Attention Mechanisms in Graph Neural Models
Any fixed GAT attention-weight matrix can be encoded as a cellular sheaf whose harmonic edges give a monotone filtration, but the framework is definitional and untested.
Discussion (0). Continue with ORCID to comment.