Article

Singular model selection for trustworthy label-free classifier evaluation

2026-08-24

Abstract excerpt

<title>Abstract</title> <p>Trustworthy deployment of a machine learning classifier requires honest uncertainty about its measured performance, including the common case where that performance is estimated without ground-truth labels by pooling several imperfect raters or models. We show that this label-free evaluation problem is singular by construction: the latent-class likelihood it relies on has a degenerate F...

Topics

Open a Topic to create a Post that cites this publication.

Identifiers and source

Literature Corpus work: 8843f692-d189-5dd1-ba50-84e84a825b10
DOI: 10.21203/rs.3.rs-10101757/v1

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

Singular model selection for trustworthy label-free classifier evaluationDOI 10.21203/rs.3.rs-10101757/v1Read

Not the Models You Are Looking For: Traditional ML Outperforms LLMs in Clinical Prediction TasksSemantic proximity Read

Achieving Explainable, Scalable, and Robust Machine Learning for Real-World ApplicationsSemantic proximity Read

Model shopping: how answer selection across AI models inflates scientific error ratesSemantic proximity Read

Quantifying and Mitigating Gender Bias in Legal Large Language Models:A Counterfactual Fairness FrameworkSemantic proximity Read

Screening Feedback for Language Models with Costly VerificationSemantic proximity Read

The Dice Roll Method: A StandardizedProtocol for Measuring Stochastic Bias in Large Language Model OutputSemantic proximity Read

Deceptive Bias Measurement in Deep Learning: Assessing Shortcut Reliance in TCGA Cancer ModelsSemantic proximity Read

The Base-Rate Trap in Generative AI Text Detection: Why Detectors Cannot Serve as Standalone Evidence of Academic Misconduct, and Who Bears the CostSemantic proximity Read

A Multi-Model, Multi-Domain Benchmark of Large Language Model Agreement with Humans and with Each Other in Sentiment ClassificationSemantic proximity Read

Cross-Fitted Contamination-Aware Generalized Empirical-Bayes Liu Shrinkage for Multinomial Logit Models under Multicollinearity and OutliersSemantic proximity Read

How LLMs Audit Each Other: Five Mechanisms of Auditor Bias in Cross-Model Peer Review Under Identity Disclosure and Cross-Lingual ConditionsSemantic proximity Read

SynthGuard-ReleaseBench: Locked-Audit Evidence for Synthetic Tabular Data ReleasesSemantic proximity Read