For simplicity, we omit the schedule coefficients in (5).

Algorithmic machine teaching studies the interaction between a teacher and a learner where the teacher selects labeled examples aiming at teaching a target hypothesis.

In this section, we provide the detailed proof of Theorem 1. We first prove the following lemma, which is a key component used in our proof.

We establish new connections between LwEQ-TD and LfS-TD by studying LwEQ-TD for different learner models based on the richness of their query functions. We show that LwEQ-TD isthesameaswc-TD[18],RTD[22,24],andNCTD[27]forahypothesis class when restricting query functions to specific families.

The difference between a representationZin IB and a statisticT(X), is that the mapping between the inputsX and the representation Z can be stochastic -- specifically a representation is a statistic of the input and independent noise X, i.e., Z = T(X,).

Now, we can exactly derive that q (G) = ˆ p( G|I) . The definitions of potential function φ and ψ follow those in JM-SGG model. Figure 1: The scene graphs generated by JM-SGG model. In these examples, factor update is able to correct some wrong relation labels ( e.g.