Improving Faithfulness in Abstractive Summarization with Contrast Candidate Generation and Selection (2104.09061v1)

Abstract: Despite significant progress in neural abstractive summarization, recent studies have shown that the current models are prone to generating summaries that are unfaithful to the original context. To address the issue, we study contrast candidate generation and selection as a model-agnostic post-processing technique to correct the extrinsic hallucinations (i.e. information not present in the source text) in unfaithful summaries. We learn a discriminative correction model by generating alternative candidate summaries where named entities and quantities in the generated summary are replaced with ones with compatible semantic types from the source document. This model is then used to select the best candidate as the final output summary. Our experiments and analysis across a number of neural summarization systems show that our proposed method is effective in identifying and correcting extrinsic hallucinations. We analyze the typical hallucination phenomenon by different types of neural summarization systems, in hope to provide insights for future work on the direction.