Open Source
Two Counterexamples for e-means
arXiv stat.ML0 views
arXiv:2611.22065v1 Announce Type: cross Abstract: e-means (Bessa et al., 2026) combines exponential averaging and cluster revival. We give two counterexamples to its published recipe with higher final loss than standard k-means (Lloyd, 1982) from identical initial centers. Mass-based revival can destroy a global optimum and make the ratio of expected losses unbounded even with k-means++ seeding.
Source:https://arxiv.org/abs/2611.22065PreviousLocally Linear Generalized Single Index RegressionNextCOSI: A Framework for Robustly Valid Inference
Related posts

Backup imutável de RDS com cofre logicamente isolado: o que PCI DSS e ISO 27001 pedem e como provar
0dev.to AWS
Show HN: NetDiag self-hosted network inventory and diagnostics
0Hacker News Show
Show HN: Map showing if GTA 6 is out in your country
0Hacker News Show
Show HN: Mtl_jobines – Montreal-inspired open world driving game
0Hacker News Show
Show HN: MR Office – open-source document, spreadsheet, slide and PDF editors
0Hacker News Show
Show HN: TaroBird – An Open-Source IM for Codex, Claude Code and Other CLIs
0Hacker News Show