AI
FACET - Facet Aware Contrastive Embeddings for Traits- Big Five Estimation from Life Narratives
Publication: cs.AI updates on arXiv.org
Author: Balakrishnan Arumugam, Priyen Parekh, Mehak Gupta, Joshua Oltmanns
Source published: 19/11/2026, 05:00:00 UTC
Briefing created: 19/11/2026, 06:21:35 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 19/11/2026, 05:03:25 UTC.
Briefings
A new framework estimates Big Five personality traits from life-narrative interviews by combining text and psycholinguistic features via contrastive embeddings. The study used 1,332 participants from the St. Louis Personality and Aging Network (SPAN) and achieved AUC up to 0.740 for facet embeddings, with domain-level R2 reaching 0.328. End-to-end training outperformed staged training, and attention pooling yielded the lowest error across all domains.
Key points
- The framework estimates Big Five personality traits from life-narrative interview text by combining text and psycholinguistic features through contrastive embeddings.
- On 1,332 participants from the St. Louis Personality and Aging Network (SPAN), facet embeddings reached an AUC of 0.740, with domain-level R2 reaching 0.328.
Context
Personality is only weakly and inconsistently signaled by written language, and existing text-based methods typically combine language-model and psycholinguistic representations at the input or decision level. This study addresses whether a more deliberate combination of these views can recover more of the signal.
Limitations
The evidence is from a single preprint (arXiv:2611.23560v1) with no peer review. Results are based on a snapshot of 1,332 participants from the St. Louis Personality and Aging Network (SPAN) study, and the framework's performance on other domains or with different data is not reported.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 19/11/2026, 06:21:35 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeiafjhcs6i3zu7nmhtu3yk6yqyg7zgxuujfigcwft4bgjopyesmgga