Latent transfer-learned features yield better overall mind wandering detection than explicit OpenFace features, and temporal models add a small improvement, especially cross-dataset¶

Claim · All claims
Evidence · 1 study · design r2 · q2
Currency · newest study 2025 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? Across all classifiers and both prediction scenarios, latent features allow better mind wandering detection than explicit features. → Bühler 2025 q2 i? When employing latent features, temporal models allow a small improvement over non-temporal classifiers, especially in cross-dataset prediction. → Bühler 2025

Evidence¶

Bühler 2025¶

Bühler, B., Bozkir, E., Goldberg, P., Sümer, Ö., D'Mello, S., Gerjets, P., Trautwein, U., Kasneci, E. (2025). From the Lab to the Wild: Examining Generalizability of Video-based Mind Wandering Detection. Int J Artif Intell Educ 35, 823–857. https://doi.org/10.1007/s40593-024-00412-2

q2 · i? · design · r2

Numerical comparison across models shown in Fig. 4 and Tables 2–3. The article reports latent features allow a better detection overall and that fine-tuning in an end-to-end CNN-LSTM leads to further improvement in cross-dataset prediction, possibly because the complete lab dataset was available for training.

"A comparison of performance along the employed feature sets over all classifiers and prediction scenarios, depicted in Fig. 4, indicates that overall the latent features allow a better mind wandering detection."

Discussion¶

Related Claims¶

• Lab-trained mind wandering detectors generalize to in-the-wild lecture viewing across task, setting, and cultural group, with BiLSTM on latent features reaching F1 = 0.352 — related
• Mind wandering detection is more accurate for female than male learners, with the gender gap larger for temporal models — related
• Transfer-learned latent facial features support above-chance video-based mind wandering detection within a lab reading dataset, with LSTM reaching F1 = 0.459 — a narrower finding that bears on this claim
• LIME analysis shows the latent features mainly encode information from eye and mouth areas of the face — related
• Optimizing the classification threshold improves cross-dataset mind wandering detection, raising person-level correlation from 0.48 to 0.8 — related