Near-Optimally Teaching the Crowd to Classify

TL;DR

This paper introduces a stochastic model for teaching classification to learners, proposes an efficient greedy algorithm called STRICT for example selection, and demonstrates its effectiveness through theoretical guarantees and experiments.

Contribution

It develops a novel stochastic learner model and a greedy teaching algorithm with provable guarantees, improving teaching efficiency in crowdsourcing and online education.

Findings

STRICT achieves exponential error reduction for linear separators.

The algorithm outperforms baseline methods in simulated and real tasks.

Theoretical analysis confirms competitiveness with optimal policies.

Abstract

How should we present training examples to learners to teach them classification rules? This is a natural problem when training workers for crowdsourcing labeling tasks, and is also motivated by challenges in data-driven online education. We propose a natural stochastic model of the learners, modeling them as randomly switching among hypotheses based on observed feedback. We then develop STRICT, an efficient algorithm for selecting examples to teach to workers. Our solution greedily maximizes a submodular surrogate objective function in order to select examples to show to the learners. We prove that our strategy is competitive with the optimal teaching policy. Moreover, for the special case of linear separators, we prove that an exponential reduction in error probability can be achieved. Our experiments on simulated workers as well as three real image annotation tasks on Amazon Mechanical Turk show the effectiveness of our teaching algorithm.