Formalizing the presumption of independence
Not on the map. Why?
Abstract (arXiv)
Mathematical proof aims to deliver confident conclusions, but a very similar process of deduction can be used to make uncertain estimates that are open to revision. A key ingredient in such reasoning is the use of a "default" estimate of mathbbE[XY] = mathbbE[X] mathbbE[Y] in the absence of any specific information about the correlation between X and Y, which we call *the presumption of independence*. Reasoning based on this heuristic is commonplace, intuitively compelling, and often quite successful -- but completely informal. In this paper we introduce the concept of a heuristic estimator as a potential formalization of this type of defeasible reasoning. We introduce a set of intuitively desirable coherence properties for heuristic estimators that are not satisfied by any existing candidates. Then we present our main open problem: is there a heuristic estimator that formalizes intuitively valid applications of the presumption of independence without also accepting spurious arguments?
Key facts. Cited by 8 papers in the map (data through 14 Oct 2026). Builds on 1 of its 16 references that are in the map. First posted on arXiv 12 Nov 2022.
Builds on
1 of 16 references.
References · 1
Built on by
Citing papers per year: 2023: 1; 2024: 4; 2025: 1; 2026: 2.
8 papers.
- 2024Compact Proofs of Model Performance via Mechanistic Interpretabilitycited by 18
- 2024Using Degeneracy in the Loss Landscape for Mechanistic Interpretabilitycited by 13
- 2023Benchmarks for Detecting Measurement Tamperingcited by 6
- 2026Towards Worst-Case Guarantees with Scale-Aware Interpretabilitycited by 3
- 2025Research Superalignment Should Advance Now with Alternating Competence and Conformity Optimizationcited by 1
- 2026Estimating the expected output of wide random MLPs more efficiently than samplingcited by 1
- 2024Polynomial Precision Dependence Solutions to Alignment Research Center Matrix Completion Problemscited by 0
- 2024Towards a Law of Iterated Expectations for Heuristic Estimatorscited by 0
- 2026Estimating the expected output of wide random MLPs more efficiently than sampling6 May 2026
- 2026Towards Worst-Case Guarantees with Scale-Aware Interpretability5 Feb 2026
- 2025Research Superalignment Should Advance Now with Alternating Competence and Conformity Optimization8 Mar 2025
- 2024Towards a Law of Iterated Expectations for Heuristic Estimators2 Oct 2024
- 2024Compact Proofs of Model Performance via Mechanistic Interpretability17 Jun 2024
- 2024Using Degeneracy in the Loss Landscape for Mechanistic Interpretability17 May 2024
- 2024Polynomial Precision Dependence Solutions to Alignment Research Center Matrix Completion Problems8 Jan 2024
- 2023Benchmarks for Detecting Measurement Tampering29 Aug 2023