CLOct 14, 2022

Self-Repetition in Abstractive Neural Summarizers

arXiv:2210.08145v1302 citationsh-index: 73
Originality Synthesis-oriented
AI Analysis

This addresses data quality issues for practitioners in text summarization, but it is incremental as it focuses on analysis rather than new methods.

The paper analyzes self-repetition in neural summarizers, finding that architectures like BART are prone to repeating content, with rates influenced by fine-tuning data, and identifies artifacts like ads in outputs.

We provide a quantitative and qualitative analysis of self-repetition in the output of neural summarizers. We measure self-repetition as the number of n-grams of length four or longer that appear in multiple outputs of the same system. We analyze the behavior of three popular architectures (BART, T5, and Pegasus), fine-tuned on five datasets. In a regression analysis, we find that the three architectures have different propensities for repeating content across output summaries for inputs, with BART being particularly prone to self-repetition. Fine-tuning on more abstractive data, and on data featuring formulaic language, is associated with a higher rate of self-repetition. In qualitative analysis we find systems produce artefacts such as ads and disclaimers unrelated to the content being summarized, as well as formulaic phrases common in the fine-tuning domain. Our approach to corpus-level analysis of self-repetition may help practitioners clean up training data for summarizers and ultimately support methods for minimizing the amount of self-repetition.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes