Work

Ambiguity, Coherence and Performance

Year: 2025

Type: article

Abstract: Imprecise Bayesianism has been proposed as an alternative to Standard Bayesianism, partly because of its tools for representing ambiguity. Instead of representing credences via precise probabilities, a set of probability distributions is used to model belief states. However, there are criticisms of Imprecise Bayesianism’s update rule. A recent alternative update rule is Alpha Cut, which evades some of the primary criticisms of Imprecise Bayesian updating. We compare Alpha Cut with Imprecise Bayesianism and another alternative update approach called Calibration. We find that Alpha Cut has problems with respect to ambiguity, coherence, and performance qualities, whereas there are more promising alternatives.

Source: Economics and Philosophy

DOI: 10.1017/s026626712510059x

Authors: William Peden, Mantas Radzvilas, Daniele Tortoli, Francesco De Pretis

Institutions: Johannes Kepler University of Linz, University of Konstanz, University of Modena and Reggio Emilia, Indiana University Bloomington

Language: English

Topic: Bayesian Modeling and Causal Inference

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

Estimated APC paid: $3,550

Funders: Università Degli Studi di Modena e Reggio Emila, Deutsche Forschungsgemeinschaft