Interpretable and Pedagogical Examples

By Smitha Milli, Pieter Abbeel, Igor Mordatch. 2017

14 citations recorded by the source corpus.

Teachers intentionally pick the most informative examples to show their students. However, if the teacher and student are neural networks, the examples that the teacher network learns to give, although effective at teaching the student, are typically uninterpretable. We show that training the student and teacher iteratively, rather than jointly, can produce interpretable teaching strategies. We evaluate interpretability by (1) measuring the similarity of the teacher's emergent strategies to intuitive strategies in each domain and (2) conducting human experiments to evaluate how effective the teacher's strategies are at teaching humans. We show that the teacher network learns to select or generate interpretable, pedagogical examples to teach rule-based, probabilistic, boolean, and hierarchical concepts.

Subfields: Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Natural Language Processing

1. Tao Qin
2. Tie-Yan Liu
3. Lijun Wu
4. Yingce Xia
5. Fei Tian
6. J. Lai
7. Yang Fan
8. Geoffrey Irving

Citations in the corpus

• AI safety via debate (2018, methodology): “Similar to Milli et al. [2017], we train the judge in advance with supervised learning, using a convolutional net that sees a sparse mask of a few pixels.”
• Learning to Teach with Dynamic Loss Functions (2018, background): “Furthermore there are research on pedagogical teaching inspired from cognitive science [44, 23, 39] in which a teacher module is responsible for providing informative examples to the learner for the sake of understanding a concept rapidly.”