SALAD: Improving Robustness and Generalization through Contrastive Learning with Structure-Aware and LLM-Driven Augmented Data
This addresses robustness issues in NLP for tasks sensitive to distribution shifts, but it is incremental as it builds on existing contrastive learning and data augmentation techniques.
The paper tackles the problem of spurious correlations in fine-tuned pre-trained language models, which harm performance on out-of-distribution data, by proposing SALAD, a method that uses structure-aware and LLM-driven augmented data for contrastive learning, resulting in improved robustness and generalization across tasks like sentiment classification, sexism detection, and natural language inference.
In various natural language processing (NLP) tasks, fine-tuning Pre-trained Language Models (PLMs) often leads to the issue of spurious correlations, which negatively impacts performance, particularly when dealing with out-of-distribution data. To address this problem, we propose SALAD}(Structure Aware and LLM-driven Augmented Data), a novel approach designed to enhance model robustness and generalization by generating structure-aware and counterfactually augmented data for contrastive learning. Our method leverages a tagging-based approach to generate structure-aware positive samples and utilizes large language models (LLMs) to generate counterfactual negative samples with diverse sentence patterns. By applying contrastive learning, SALAD enables the model to focus on learning the structural relationships between key sentence components while minimizing reliance on spurious correlations. We validate our approach through experiments on three tasks: Sentiment Classification, Sexism Detection, and Natural Language Inference. The results demonstrate that SALAD not only improves model robustness and performance across different environments but also enhances generalization to out-of-distribution datasets and cross-domain scenarios.