Thread follow-up: the tail-area criterion traces back to 1990s Bayesian literature
BlackHC · x · 2026-09-09
The author adds scholarly lineage: the tail-area criterion has a clear history — the CLML gap between two models is a log partial Bayes factor. A 1990s trade-off transfers directly: a fixed-fraction tail keeps only a constant signal, so consistency can fail unless k/N → 1 (O'Hagan 1995; Mukhopadhyay, Ghosh & Berger 2005).
Further reading: Dawid 1984 (prequential assessment), Fong & Holmes 2020 (evidence as cumulative cross-validation), Lyle et al. 2020 and Ru et al. 2021 (training speed as a selection signal), and Watanabe on learning coefficients for singular models like neural networks.
More from Research
- Open-source decision models sold their own base in Red Alert 2 — then learned until no human could beat them — heathercmiller · 2026-10-27
- DeepMind paper: LLMs can't sample randomly on their own — pass random numbers from code instead — rohanpaul_ai · 2026-10-27
- World Model Assigns Weight and Friction to Objects, Then Simulates with a Physics Engine — jonstephens85 · 2026-10-27
- Token Collapse: Study Finds Graph Soft Prompting Tokens in LLMs Fail to Encode Structure — Sapienza · 2026-10-27
- Mathathon 2.0 launches with Terence Tao's SAIR, $1.5M in compute for teams — _sathvikr · 2026-10-27
- The 'sweet lesson' is real: two kinds of ML problems need opposite strategies — mike64_t · 2026-10-27