Auto-GDA: Automatic Domain Adaptation for Efficient Grounding Verification in Retrieval-Augmented Generation

TL;DR

Auto-GDA introduces an unsupervised, iterative domain adaptation framework that enhances lightweight NLI models for more efficient grounding verification in RAG systems, reducing computational costs while maintaining high accuracy.

Contribution

It proposes a novel automatic generative domain adaptation method that improves NLI models for RAG without requiring labeled data or extensive fine-tuning.

Findings

Synthetic data generated by Auto-GDA improves NLI model performance.

Models fine-tuned on Auto-GDA data outperform teacher models.

Auto-GDA achieves comparable results to large LLMs at 10% of their computational cost.

Abstract

While retrieval-augmented generation (RAG) has been shown to enhance factuality of large language model (LLM) outputs, LLMs still suffer from hallucination, generating incorrect or irrelevant information. A common detection strategy involves prompting the LLM again to assess whether its response is grounded in the retrieved evidence, but this approach is costly. Alternatively, lightweight natural language inference (NLI) models for efficient grounding verification can be used at inference time. While existing pre-trained NLI models offer potential solutions, their performance remains subpar compared to larger models on realistic RAG inputs. RAG inputs are more complex than most datasets used for training NLI models and have characteristics specific to the underlying knowledge base, requiring adaptation of the NLI models to a specific target domain. Additionally, the lack of labeled instances in the target domain makes supervised domain adaptation, e.g., through fine-tuning, infeasible. To address these challenges, we introduce Automatic Generative Domain Adaptation (Auto-GDA). Our framework enables unsupervised domain adaptation through synthetic data generation. Unlike previous methods that rely on handcrafted filtering and augmentation strategies, Auto-GDA employs an iterative process to continuously improve the quality of generated samples using weak labels from less efficient teacher models and discrete optimization to select the most promising augmented samples. Experimental results demonstrate the effectiveness of our approach, with models fine-tuned on synthetic data using Auto-GDA often surpassing the performance of the teacher model and reaching the performance level of LLMs at 10% of their computational cost.