Automatic Detection of Reflective Thinking in Mathematical Problem Solving based on Unconstrained Bodily Exploration

TL;DR

This study introduces a dataset and machine learning methods to automatically detect moments of reflective thinking in children during mathematical problem solving through analysis of body movement data.

Contribution

The paper presents the weDraw-1 Movement Dataset and demonstrates effective neural network-based detection of reflective thinking from unconstrained bodily exploration.

Findings

Achieved 0.73 F1 score for detecting reflective thinking periods.

Achieved 0.79 F1 score for end-to-end detection from raw sensor data.

Detected short reflective segments as brief as 4 seconds with 0.64 F1 score.

Abstract

For technology (like serious games) that aims to deliver interactive learning, it is important to address relevant mental experiences such as reflective thinking during problem solving. To facilitate research in this direction, we present the weDraw-1 Movement Dataset of body movement sensor data and reflective thinking labels for 26 children solving mathematical problems in unconstrained settings where the body (full or parts) was required to explore these problems. Further, we provide qualitative analysis of behaviours that observers used in identifying reflective thinking moments in these sessions. The body movement cues from our compilation informed features that lead to average F1 score of 0.73 for automatic detection of reflective thinking based on Long Short-Term Memory neural networks. We further obtained 0.79 average F1 score for end-to-end detection of reflective thinking periods, i.e. based on raw sensor data. Finally, the algorithms resulted in 0.64 average F1 score for period subsegments as short as 4 seconds. Overall, our results show the possibility of detecting reflective thinking moments from body movement behaviours of a child exploring mathematical concepts bodily, such as within serious game play.