aleatoric-uncertainty

Tag

Cards List
#aleatoric-uncertainty

Verifiable Rewards for Calibrated Probabilistic Forecasting

arXiv cs.LG · 19h ago Cached

The paper proposes a verifiable label-free reward for training calibrated probabilistic forecasters using reinforcement learning, avoiding the calibration degradation that occurs when rewarding single outcomes. Applied to NFL win probability, a 7B model trained with this reward achieves calibration comparable to the betting market.

0 favorites 0 likes
← Back to home

Submit Feedback