AI & ML
Transport-Free Lower Bounds for Ollivier--Ricci Curvature in Higher-Order Networks: Scope, Tightness, and Computational Applications
arXiv stat.ML0 views
arXiv:2405.13302v2 Announce Type: replace Abstract: We study how the Bauer--Jost--Liu curvature bound applies to hypergraph Ollivier--Ricci curvature (ORC) and how its accuracy depends on local mass and geometry. An explicit weighted-graph representation shows how the graph theorem applies. A transport construction transfers the bounds to monotone pairwise hyperedge aggregations. We analyze endpoint cancellation under laziness, quantify the exact improvement over the earlier mixture bound, and characterize the extremal geometry at fixed endpoint and overlap statistics. Further results quantify worst-case slack, stability, and exactness on isolated hyperedges. A capacity-constrained short-detour refinement uses additional geometric information and improves monotonically as the query budget increases. As a computational application, a threshold cascade preserves exact transport decisions. Numerical verification and public-hypergraph experiments examine tightness, walk dependence, and aggregation, including evaluations on 42,940 finite pairwise comparisons and 64,620 public-data instances. Together, the results specify the scope, equality mechanisms, and computational uses of BJL-type hypergraph bounds.
Source:https://arxiv.org/abs/2405.13302PreviousGradient Descent on Two ReLU Neurons: Global Landscape and Bifurcation DynamicsNextCritically-Damped Higher-Order Langevin Dynamics for Generative Modeling
Related posts
ISC Stormcast For Friday, November 6th, 2026 https://isc.sans.edu/podcastdetail/10168, (Fri, Nov 6th)
0SANS ISC
Astra did 40% less safety work — OpenAI's own report
0dev.to Discuss

The Facilitator's Stack: A Checklist From Invitation to Follow-Up
0dev.to Productivity
Wabi Not for You? Pick the Right Alternative by What You Actually Need
0dev.to Productivity

RAG vs fine-tuning: which should your LLM app use?
0dev.to Machine Learning
When two independent model panels both say an MMLU answer key is wrong, experts found an error 1 time in 3
0dev.to Machine Learning