Debiased Inference for AI-Generated Data without Gold-Standard Labels: Identification via Multiple Imperfect Measurements
stat.ME, cs.AI, cs.CL, cs.LG, stat.ML
Submitted: 2026-08-18
Updated: 2026-09-01
Terminology
Sources
- PPI++: Efficient Prediction-Powered Inference
- Inference for Regression with Variables Generated by AI or Machine Learning
- Large Language Model Hacking: Quantifying the Hidden Risks of Using LLMs for Text Annotation
- A Unifying Framework for Robust and Efficient Inference with Unstructured Data
- Partial Identification from LLM Prompts
- Proximal Causal Inference for Hidden Outcomes
- Correlated Errors in Large Language Models
- Surrogate Representation Inference for Text and Image Annotations
- Benchmark Illusion: Disagreement among LLMs and Its Scientific Consequences
- Causal Inference with a Hidden Treatment
Related papers
- Doubly robust inference via calibration
- Bayesian Empirical Bayes: Simultaneous Inference from Probabilistic Symmetries
- Weak-Form Recovery of Stochastic Generators
- Optimal Watermark Localization in Mixed-Source Large Language Model Texts
- Optimized Conformal Selection: Powerful Selective Inference After Data-Dependent Score Construction
- Small area learning