Can vectors read minds better than experts? Comparing data augmentation strategies for the automated scoring of children's mindreading ability

TL;DR

This study compares seven data augmentation strategies for automatically scoring children's mindreading ability, demonstrating that task-specific augmentations improve performance and generalization, with vector-based methods performing poorly.

Contribution

The paper introduces and evaluates multiple data augmentation strategies for mindreading assessment, establishing a new state-of-the-art and providing insights into their effectiveness.

Findings

Task-specific augmentations outperform task-agnostic ones.

Automatic vector-based augmentations perform poorly.

Augmentation improves generalization to unseen data.

Abstract

In this paper we implement and compare 7 different data augmentation strategies for the task of automatic scoring of children's ability to understand others' thoughts, feelings, and desires (or "mindreading"). We recruit in-domain experts to re-annotate augmented samples and determine to what extent each strategy preserves the original rating. We also carry out multiple experiments to measure how much each augmentation strategy improves the performance of automatic scoring systems. To determine the capabilities of automatic systems to generalize to unseen data, we create UK-MIND-20 - a new corpus of children's performance on tests of mindreading, consisting of 10,320 question-answer pairs. We obtain a new state-of-the-art performance on the MIND-CA corpus, improving macro-F1-score by 6 points. Results indicate that both the number of training examples and the quality of the augmentation strategies affect the performance of the systems. The task-specific augmentations generally outperform task-agnostic augmentations. Automatic augmentations based on vectors (GloVe, FastText) perform the worst. We find that systems trained on MIND-CA generalize well to UK-MIND-20. We demonstrate that data augmentation strategies also improve the performance on unseen data.