Institution profile

U.S. Department of Veterans Affairs

Academic institutionnorthamerica · us
Official website
Research library1linked papers
Opportunities0open roles
Selected work

Representative Papers

Advancing Minority Stress Detection with Transformers: Insights from the Social Media Datasets

Sep 02, 2025

This study addresses the challenge of detecting latent minority stress—experienced by sexual and gender minority (SGM) individuals on social media—through a graph-enhanced Transformer framework. Methodologically, it integrates multiple pretrained language models (ELECTRA, BERT, RoBERTa, BART), couples them with graph neural networks to explicitly model user interactions and conversational relationships, and systematically evaluates supervised fine-tuning, zero-shot, and few-shot learning on highly imbalanced Reddit data. Results demonstrate, for the first time empirically, that explicit graph-structured encoding of social context significantly improves minority stress detection accuracy—particularly for fine-grained linguistic markers such as identity concealment, internalized stigma, and support-seeking behavior. Supervised fine-tuning consistently outperforms zero-shot and few-shot alternatives. The work advances digital mental health interventions by delivering an interpretable, deployable computational paradigm grounded in both linguistic and relational modeling.

0 citationsRead paper
Recent publications

Latest Papers

Advancing Minority Stress Detection with Transformers: Insights from the Social Media Datasets

Sep 02, 2025

This study addresses the challenge of detecting latent minority stress—experienced by sexual and gender minority (SGM) individuals on social media—through a graph-enhanced Transformer framework. Methodologically, it integrates multiple pretrained language models (ELECTRA, BERT, RoBERTa, BART), couples them with graph neural networks to explicitly model user interactions and conversational relationships, and systematically evaluates supervised fine-tuning, zero-shot, and few-shot learning on highly imbalanced Reddit data. Results demonstrate, for the first time empirically, that explicit graph-structured encoding of social context significantly improves minority stress detection accuracy—particularly for fine-grained linguistic markers such as identity concealment, internalized stigma, and support-seeking behavior. Supervised fine-tuning consistently outperforms zero-shot and few-shot alternatives. The work advances digital mental health interventions by delivering an interpretable, deployable computational paradigm grounded in both linguistic and relational modeling.

0 citationsRead paper