Multi-task PPI framework uses cross-task recalibration to improve inference power across related tasks, with a proof that gains require nonlinear proxy-ground-truth structure, shown on synthetic data and a 2024 election LM audit case study.
Another look at inference after prediction
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
verdicts
UNVERDICTED 3representative citing papers
The MLA-UCB algorithm uses ML-generated surrogate rewards from auxiliary data to provably lower cumulative regret in multi-armed bandits, achieving asymptotic optimality under joint Gaussian assumptions without requiring knowledge of the true-surrogate covariance.
Adaptive Matrix Validation calibrates AI-mapped survey responses using sparse randomized validation questions from other respondents then corrects with the target's own answers, with estimators and planning formulas for means, subgroups, and regressions.
citing papers explorer
-
Prediction-Powered Inference Across Many Tasks for AI Evaluation & Social Science Research
Multi-task PPI framework uses cross-task recalibration to improve inference power across related tasks, with a proof that gains require nonlinear proxy-ground-truth structure, shown on synthetic data and a 2024 election LM audit case study.
-
Multi-Armed Bandits With Machine Learning-Generated Surrogate Rewards
The MLA-UCB algorithm uses ML-generated surrogate rewards from auxiliary data to provably lower cumulative regret in multi-armed bandits, achieving asymptotic optimality under joint Gaussian assumptions without requiring knowledge of the true-surrogate covariance.
-
When Surveys Become Conversations: Adaptive Matrix Validation for AI-Assisted Interviews
Adaptive Matrix Validation calibrates AI-mapped survey responses using sparse randomized validation questions from other respondents then corrects with the target's own answers, with estimators and planning formulas for means, subgroups, and regressions.