Learning from Generalization Patterns: An Evaluation-Driven Approach to Enhanced Data Augmentation for Fine-Tuning Small Language Models
This work addresses the challenge of enhancing SLM accuracy for domain-specific tasks with reduced manual effort, though it appears incremental as it builds on existing data augmentation techniques.
The paper tackles the performance gap of small language models (SLMs) on complex tasks by introducing PaDA-Agent, an evaluation-driven data augmentation approach that discovers failure patterns from validation data to reduce generalization errors, resulting in significant improvements over state-of-the-art LLM-based methods for fine-tuning the Llama 3.2 1B Instruct model.
Small Language Models (SLMs) offer compelling advantages in deployment cost and latency, but their accuracy often lags behind larger models, particularly for complex domain-specific tasks. While supervised fine-tuning can help bridge this performance gap, it requires substantial manual effort in data preparation and iterative optimization. We present PaDA-Agent (Pattern-guided Data Augmentation Agent), an evaluation-driven approach that streamlines the data augmentation process for SLMs through coordinated operations. Unlike state-of-the-art approaches that focus on model training errors only and generating error-correcting samples, PaDA-Agent discovers failure patterns from the validation data via evaluations and drafts targeted data augmentation strategies aiming to directly reduce the generalization gap. Our experimental results demonstrate significant improvements over state-of-the-art LLM-based data augmentation approaches for Llama 3.2 1B Instruct model fine-tuning.