Definition
Training one system from raw input to final answer in one shot, instead of hand-engineering each step in between.
A training paradigm in which a single model is optimized directly on a downstream objective, replacing hand-crafted feature pipelines with learned representations.