NeFut Logo NeFut
Admin Login

[CS.AI] Different Representation Learning Objectives Recover Distinct Latent Structures from the Same Psychometric Data

Published at: 2026-09-02 22:00 Last updated: 2026-09-03 02:56
#Machine Learning #optimization #Neural

Psychometric questionnaires contain rich item‑level information, yet it remains unclear whether different representation learning objectives recover the same latent organization. We investigated this question using 757 matched teacher‑child pairs from the baseline assessment of the Cyprus ProW preschool trial. Child responses on the SDQ, ASBI, and CBRS were analyzed with principal component analysis (PCA) and clustering, yielding four behavioral phenotypes. A contrastive learning objective substantially outperformed PCA‑based representations in a teacher‑child retrieval task, raising Top‑1 accuracy from 0.13% to 7.27% and Top‑10 accuracy from 1.98% to 56.14%. However, contrastive representations preserved the behavioral phenotype structure less effectively than PCA. A multi‑task objective that jointly optimizes alignment and behavioral prediction partially restored the phenotype organization but reduced retrieval performance. These findings indicate that teacher‑child correspondence and behavioral phenotypes constitute distinct forms of latent organization, and that the latent structure recovered from linked psychometric data depends on the chosen representation learning objective.

Review

Original Source: https://arxiv.org/abs/2609.00100

[h] Back to Home