K-COMP: Retrieval-Augmented Medical Domain Question Answering With Knowledge-Injected Compressor
This addresses accuracy and trust issues in medical QA systems, but it is incremental as it builds on existing retrieval-augmented methods.
The paper tackles the problem of retrieval-augmented question answering in medical domains, where retrieved documents are long and contain irrelevant or inaccurate information, leading to model hallucinations. The result is K-comp, a knowledge-injected compressor that generates prior knowledge and compresses passages to improve accuracy, though no concrete numbers are provided.
Retrieval-augmented question answering (QA) integrates external information and thereby increases the QA accuracy of reader models that lack domain knowledge. However, documents retrieved for closed domains require high expertise, so the reader model may have difficulty fully comprehending the text. Moreover, the retrieved documents contain thousands of tokens, some unrelated to the question. As a result, the documents include some inaccurate information, which could lead the reader model to mistrust the passages and could result in hallucinations. To solve these problems, we propose K-comp (Knowledge-injected compressor) which provides the knowledge required to answer correctly. The compressor automatically generates the prior knowledge necessary to facilitate the answer process prior to compression of the retrieved passages. Subsequently, the passages are compressed autoregressively, with the generated knowledge being integrated into the compression process. This process ensures alignment between the question intent and the compressed context. By augmenting this prior knowledge and concise context, the reader models are guided toward relevant answers and trust the context.