Why So Gullible? Enhancing the Robustness of Retrieval-Augmented Models against Counterfactual Noise
Giwon Hong, Jeonghwan Kim, Junmo Kang, Sung-Hyon Myaeng, Joyce Jiyoung Whang · 2024
Most existing retrieval-augmented language models (LMs) assume a naïve dichotomy within a retrieved document set: queryrelevance and irrelevance.Our work investigates a more challenging scenario in which even the "relevant" documents may contain misleading or incorrect information, causing conflict among the retrieved documents and thereby negatively influencing model decisions as noise.We observe that existing LMs are highly brittle to the presence of conflicting information in both the fine-tuning and incontext few-shot learning scenarios.We propose approaches for handling knowledge conflicts among retrieved documents by explicitly fine-tuning a discriminator or prompting GPT-3.5 to elicit its discriminative capability.Our empirical results on open-domain QA show that these approaches significantly enhance model robustness.We also provide our findings on incorporating the fine-tuned discriminator's decision into the in-context learning process, proposing a way to exploit the benefits of two disparate learning schemes.Alongside our findings, we provide MACNOISE, a machine-generated, conflict-induced dataset to further encourage research in this direction 1 .