Stochastically Dominant Peer Prediction
This work addresses the challenge of incentivizing truthful human feedback for machine learning tasks like learning from noisy labels and AI alignment, offering a stronger guarantee than traditional mechanisms, though it is incremental in improving sensitivity and applicability.
The paper tackles the problem of eliciting reliable human feedback in peer prediction mechanisms by proposing stochastically dominant truthfulness (SD-truthfulness), which ensures truthful reporting for a wide range of monotone utility functions, and introduces an enforced agreement (EA) mechanism that achieves the highest sensitivity among known SD-truthful mechanisms in binary-signal settings.
Eliciting reliable human feedback is essential for many machine learning tasks, such as learning from noisy labels and aligning AI systems with human preferences. Peer prediction mechanisms incentivize truthful reporting without ground truth verification by scoring agents based on correlations with peers. Traditional mechanisms, which ensure that truth-telling maximizes the expected scores in equilibrium, can elicit honest information while assuming agents' utilities are linear functions of their scores. However, in practice, non-linear payment rules are usually preferred, or agents' utilities are inherently non-linear. We propose stochastically dominant truthfulness (SD-truthfulness) as a stronger guarantee: the score distribution of truth-telling stochastically dominates all other strategies, incentivizing truthful reporting for a wide range of monotone utility functions. Our first observation is that no existing peer prediction mechanism naturally satisfies this criterion without strong assumptions. A simple solution -- rounding scores into binary lotteries -- can enforce SD-truthfulness, but often degrades sensitivity, a key property related to fairness and statistical efficiency. We demonstrate how a more careful application of rounding can better preserve sensitivity. Furthermore, we introduce a new enforced agreement (EA) mechanism that is theoretically guaranteed to be SD-truthful in binary-signal settings under mild assumptions, and empirically achieves the highest sensitivity among all known SD-truthful mechanisms.