{"paper":{"title":"Sheaf Neural Networks on SPD Manifolds: Second-Order Geometric Representation Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Sheaf neural networks defined on SPD manifolds represent second-order geometric features that Euclidean sheaves cannot.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Anna Wienhard, Ce Ju, Diaaeldin Taha, Hao Li, Huitao Feng, Junwen Dong, Kelin Xia, Yuhan Peng, Yuzhi Zeng","submitted_at":"2026-04-22T08:09:04Z","abstract_excerpt":"Graph neural networks face two fundamental challenges rooted in the linear structure of Euclidean vector spaces: (1) Current architectures represent geometry through vectors (directions, gradients), yet many tasks require matrix-valued representations that capture relationships between directions-such as how atomic orientations covary in a molecule. These second-order representations are naturally captured by points on the symmetric positive definite matrices (SPD) manifold; (2) Standard message passing applies shared transformations across edges. Sheaf neural networks address this via edge-sp"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"SPD-valued sheaves are strictly more expressive than Euclidean sheaves: they admit consistent configurations (global sections) that vector-valued sheaves cannot represent, directly translating to richer learned representations.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The SPD manifold admits a Lie group structure, enabling well-posed analogs of sheaf operators without projecting to Euclidean space.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Sheaf neural networks on the SPD manifold enable strictly more expressive second-order geometric representations than Euclidean versions and achieve SOTA results on most MoleculeNet benchmarks.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Sheaf neural networks defined on SPD manifolds represent second-order geometric features that Euclidean sheaves cannot.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"5dabebeeec228c7affbf9d0cf4c3a722a864eec41291f3d1af6014d3b209ddbf"},"source":{"id":"2604.20308","kind":"arxiv","version":2},"verdict":{"id":"bce2194c-75a9-4d08-8ebc-3d4555642811","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T00:35:29.506272Z","strongest_claim":"SPD-valued sheaves are strictly more expressive than Euclidean sheaves: they admit consistent configurations (global sections) that vector-valued sheaves cannot represent, directly translating to richer learned representations.","one_line_summary":"Sheaf neural networks on the SPD manifold enable strictly more expressive second-order geometric representations than Euclidean versions and achieve SOTA results on most MoleculeNet benchmarks.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The SPD manifold admits a Lie group structure, enabling well-posed analogs of sheaf operators without projecting to Euclidean space.","pith_extraction_headline":"Sheaf neural networks defined on SPD manifolds represent second-order geometric features that Euclidean sheaves cannot."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.20308/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-21T14:43:33.126206Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-20T02:02:28.303246Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"08710cbd797c96aa848ef86e29f75cd7f391dec743c2f381b128d454162f2873"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}