Topic channel
Information
1 paper across 2 private calibration editions.
- 01Training AI for When Humans Will Use It
A theoretical model argues that training an AI solely for unconditional predictive accuracy can be suboptimal once human verification and downstream decisions are considered.