Learning Theory2026-11-13learning‑theorypac‑learningsample‑complexityvc‑dimensionshatteringmistake‑boundshalvinginductive‑biasgeneralisationcomp9417agenticHow many examples a learner needs, and how many mistakes it makes, as a function of the size or VC dimension of its hypothesis space: PAC learning, Haussler's bound, shattering and mistake bounds.