Figure 7.

The flow chart of general procedures in the adaptive learning system: (1) the knowledge state $\mathit{s}(t)$ is partially observed by responses, estimated as $\hat{\mathit{s}}(t)$ in the assessment model; (2) given $\hat{\mathit{s}}(t)$, action $a(t)$ is recommended following the policy $\pi$; and (3) according to the learning action $a(t)$, the learning model determines the next knowledge state $\mathit{s}(t+1)$.

Note. IRT = item response theory.