Iterative Teaching by Data Hallucination

TL;DR

This paper introduces Data Hallucination Teaching (DHT), a novel iterative machine teaching method that generates input data in continuous spaces to improve teaching effectiveness for various learners.

Contribution

The paper proposes a new data hallucination approach for iterative teaching in continuous spaces, enhancing the teacher's ability to generate effective training examples.

Findings

DHT outperforms traditional methods in various teaching scenarios.

Empirical results demonstrate the effectiveness of data hallucination in improving learner performance.

The approach is applicable to both linear and neural learners in different settings.

Abstract

We consider the problem of iterative machine teaching, where a teacher sequentially provides examples based on the status of a learner under a discrete input space (i.e., a pool of finite samples), which greatly limits the teacher's capability. To address this issue, we study iterative teaching under a continuous input space where the input example (i.e., image) can be either generated by solving an optimization problem or drawn directly from a continuous distribution. Specifically, we propose data hallucination teaching (DHT) where the teacher can generate input data intelligently based on labels, the learner's status and the target concept. We study a number of challenging teaching setups (e.g., linear/neural learners in omniscient and black-box settings). Extensive empirical results verify the effectiveness of DHT.