Attention-Based LSTM for Psychological Stress Detection from Spoken Language Using Distant Supervision

Abstract

We propose a Long Short-Term Memory (LSTM) with attention mechanism to classify psychological stress from self-conducted interview transcriptions. We apply distant supervision by automatically labeling tweets based on their hashtag content, which complements and expands the size of our corpus. This additional data is used to initialize the model parameters, and which it is fine-tuned using the interview data. This improves the model's robustness, especially by expanding the vocabulary size. The bidirectional LSTM model with attention is found to be the best model in terms of accuracy (74.1%) and f-score (74.3%). Furthermore, we show that distant supervision fine-tuning enhances the model's performance by 1.6% accuracy and 2.1% f-score. The attention mechanism helps the model to select informative words.

arXiv commentAccepted in ICASSP 2018

References in corpus (1)

• Distributed Representations of Words and Phrases and their Compositionality

Cited by in corpus (8)

• Calibration of Transformer-based Models for Identifying Stress and Depression in Social Media
• Progress Notes Classification and Keyword Extraction using Attention-based Deep Learning Models with BERT
• Emo2Vec: Learning Generalized Emotion Representation by Multi-task Training
• Nora: The Well-Being Coach
• Clickbait? Sensational Headline Generation with Auto-tuned Reinforcement Learning
• Autoencoders as Tools for Program Synthesis
• Dreaddit: A Reddit Dataset for Stress Analysis in Social Media
• Interpreting Criminal Charge Prediction and Its Algorithmic Bias via Quantum-Inspired Complex Valued Networks