MACSUM: Controllable Summarization with Mixed Attributes

Abstract

Controllable summarization allows users to generate customized summaries with specified attributes. However, due to the lack of designated annotations of controlled summaries, existing work has to craft pseudo datasets by adapting generic summarization benchmarks. Furthermore, most research focuses on con¬trolling single attributes individually (e.g., a short summary or a highly abstractive summary) rather than controlling a mix of attributes together (e.g., a short and highly abstractive summary). In this paper, we propose MAC-SUM, the first human-annotated summariza¬tion dataset for controlling mixed attributes. It contains source texts from two domains, news articles and dialogues, with human-annotated summaries controlled by five designed at-tributes (Length, Extractiveness, Specificity, Topic, and Speaker). We propose two simple and effective parameter-efficient approaches for the new task of mixed controllable sum-marization based on hard prompt tuning and soft prefix tuning. Results and analysis demon¬strate that hard prompt models yield the best performance on most metrics and human eval¬uations. However, mixed-attribute control is still challenging for summarization tasks. Our dataset and code are available at https://github.com/psunlpgroup/MACSum.