Teaching with Limited Information on the Learner’s Behaviour

Ferdinando Cicalese, Sergio Filho, Eduardo Laber, Marco Molinaro
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:2016-2026, 2020.

Abstract

Machine Teaching studies how efficiently a Teacher can guide a Learner to a target hypothesis. We focus on the model of Machine Teaching with a black box learner introduced in [Dasgupta et al., ICML 2019], where the teaching is done interactively without having any knowledge of the Learner’s algorithm and class of hypotheses, apart from the fact that it contains the target hypothesis $h^*$. We first refine some existing results for this model and, then, we study new variants of it. Motivated by the realistic possibility that $h^*$ is not available to the learner, we consider the case where the teacher can only aim at having the learner converge to a best available approximation of $h^*$. We also consider weaker black box learners, where, in each round, the choice of the consistent hypothesis returned to the Teacher is not adversarial, and in particular, we show that better provable bounds can be obtained for a type of Learner that moves to the next hypothesis smoothly, preferring hypotheses that are close to the current one; and for another type of Learner that can provide to the Teacher hypotheses chosen at random among those consistent with the examples received so far. Finally, we present an empirical evaluation of our basic interactive teacher on real datasets.

Cite this Paper


BibTeX
@InProceedings{pmlr-v119-cicalese20a, title = {Teaching with Limited Information on the Learner’s Behaviour}, author = {Cicalese, Ferdinando and Filho, Sergio and Laber, Eduardo and Molinaro, Marco}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {2016--2026}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/cicalese20a/cicalese20a.pdf}, url = {https://proceedings.mlr.press/v119/cicalese20a.html}, abstract = {Machine Teaching studies how efficiently a Teacher can guide a Learner to a target hypothesis. We focus on the model of Machine Teaching with a black box learner introduced in [Dasgupta et al., ICML 2019], where the teaching is done interactively without having any knowledge of the Learner’s algorithm and class of hypotheses, apart from the fact that it contains the target hypothesis $h^*$. We first refine some existing results for this model and, then, we study new variants of it. Motivated by the realistic possibility that $h^*$ is not available to the learner, we consider the case where the teacher can only aim at having the learner converge to a best available approximation of $h^*$. We also consider weaker black box learners, where, in each round, the choice of the consistent hypothesis returned to the Teacher is not adversarial, and in particular, we show that better provable bounds can be obtained for a type of Learner that moves to the next hypothesis smoothly, preferring hypotheses that are close to the current one; and for another type of Learner that can provide to the Teacher hypotheses chosen at random among those consistent with the examples received so far. Finally, we present an empirical evaluation of our basic interactive teacher on real datasets.} }
Endnote
%0 Conference Paper %T Teaching with Limited Information on the Learner’s Behaviour %A Ferdinando Cicalese %A Sergio Filho %A Eduardo Laber %A Marco Molinaro %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-cicalese20a %I PMLR %P 2016--2026 %U https://proceedings.mlr.press/v119/cicalese20a.html %V 119 %X Machine Teaching studies how efficiently a Teacher can guide a Learner to a target hypothesis. We focus on the model of Machine Teaching with a black box learner introduced in [Dasgupta et al., ICML 2019], where the teaching is done interactively without having any knowledge of the Learner’s algorithm and class of hypotheses, apart from the fact that it contains the target hypothesis $h^*$. We first refine some existing results for this model and, then, we study new variants of it. Motivated by the realistic possibility that $h^*$ is not available to the learner, we consider the case where the teacher can only aim at having the learner converge to a best available approximation of $h^*$. We also consider weaker black box learners, where, in each round, the choice of the consistent hypothesis returned to the Teacher is not adversarial, and in particular, we show that better provable bounds can be obtained for a type of Learner that moves to the next hypothesis smoothly, preferring hypotheses that are close to the current one; and for another type of Learner that can provide to the Teacher hypotheses chosen at random among those consistent with the examples received so far. Finally, we present an empirical evaluation of our basic interactive teacher on real datasets.
APA
Cicalese, F., Filho, S., Laber, E. & Molinaro, M.. (2020). Teaching with Limited Information on the Learner’s Behaviour. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:2016-2026 Available from https://proceedings.mlr.press/v119/cicalese20a.html.

Related Material