A Novel Approach to Train Diverse Types of Language Models for Health Mention Classification of Tweets
This work addresses a domain-specific problem in NLP for health informatics by improving disease detection in social media data, though it is incremental as it builds on existing adversarial training techniques.
The paper tackles the challenge of health mention classification in tweets, where non-health and figurative uses of disease words complicate detection, by proposing a novel training approach using adversarial training with Gaussian noise perturbation at different transformer layers and contrastive loss. Results show significant performance improvements over baseline methods on the PHM2017 dataset, with analysis indicating that noise addition at earlier and final layers enhances performance, while intermediate layers degrade it.
Health mention classification deals with the disease detection in a given text containing disease words. However, non-health and figurative use of disease words adds challenges to the task. Recently, adversarial training acting as a means of regularization has gained popularity in many NLP tasks. In this paper, we propose a novel approach to train language models for health mention classification of tweets that involves adversarial training. We generate adversarial examples by adding perturbation to the representations of transformer models for tweet examples at various levels using Gaussian noise. Further, we employ contrastive loss as an additional objective function. We evaluate the proposed method on the PHM2017 dataset extended version. Results show that our proposed approach improves the performance of classifier significantly over the baseline methods. Moreover, our analysis shows that adding noise at earlier layers improves models' performance whereas adding noise at intermediate layers deteriorates models' performance. Finally, adding noise towards the final layers performs better than the middle layers noise addition.