Can LMs Generalize to Future Data? An Empirical Analysis on Text Summarization

TL;DR

This paper investigates how well pre-trained language models generalize to future data in text summarization, revealing that memorized knowledge impacts faithfulness and current methods struggle to improve future performance.

Contribution

The study introduces TempoSum, a new benchmark for evaluating temporal generalization in summarization models, and provides insights into the limitations of existing faithfulness enhancement techniques.

Findings

Parametric knowledge affects summary faithfulness on future data

Existing faithfulness methods do not reliably improve future performance

Models struggle to generalize to data from 2010 to 2022

Abstract

Recent pre-trained language models (PLMs) achieve promising results in existing abstractive summarization datasets. However, existing summarization benchmarks overlap in time with the standard pre-training corpora and finetuning datasets. Hence, the strong performance of PLMs may rely on the parametric knowledge that is memorized during pre-training and fine-tuning. Moreover, the knowledge memorized by PLMs may quickly become outdated, which affects the generalization performance of PLMs on future data. In this work, we propose TempoSum, a novel benchmark that contains data samples from 2010 to 2022, to understand the temporal generalization ability of abstractive summarization models. Through extensive human evaluation, we show that parametric knowledge stored in summarization models significantly affects the faithfulness of the generated summaries on future data. Moreover, existing faithfulness enhancement methods cannot reliably improve the faithfulness of summarization models on future data. Finally, we discuss several recommendations to the research community on how to evaluate and improve the temporal generalization capability of text summarization models.