This Email Could Save Your Life: Introducing the Task of Email Subject Line Generation

Abstract

Given the overwhelming number of emails, an effective subject line becomes essential to better inform the recipient of the email's content. In this paper, we propose and study the task of email subject line generation: automatically generating an email subject line from the email body. We create the first dataset for this task and find that email subject line generation favor extremely abstractive summary which differentiates it from news headline generation or news single document summarization. We then develop a novel deep learning method and compare it to several baselines as well as recent state-of-the-art text summarization systems. We also investigate the efficacy of several automatic metrics based on correlations with human judgments and propose a new automatic evaluation metric. Our system outperforms competitive baselines given both automatic and human evaluations. To our knowledge, this is the first work to tackle the problem of effective email subject line generation.

AdaptSum: Towards Low-Resource Domain Adaptation for Abstractive Summarization

“…Email Zhang and Tetreault (2019) introduced an abstractive business and personal email summarization dataset which consists of email and subject pairs. We collect the unlabeled email corpus from the Enron Email Dataset.…”

“…Yet, despite their practicality, very few studies have used domain adaptation methods on the lowresource scenario for the abstractive summarization task. To address this research gap, we present AdaptSum, the first benchmark to simulate the low-resource domain Adaptation setting for abstractive Summarization systems with a combination of existing datasets across six diverse domains (dialog (Gliwa et al, 2019), email (Zhang and Tetreault, 2019), movie review (Wang and Ling, 2016), debate (Wang and Ling, 2016), social media (Kim et al, 2019), and science (Yasunaga et al, 2019)), and for each domain, we reduce the number of training samples to a small quantity so as to create a low-resource scenario.…”

“…Experimental SetupTraining Details We evaluate all of our models on AdaptSum. For the dialog and email domains, we use the standard splits of(Gliwa et al, 2019;Zhang and Tetreault, 2019), while for movie review, debate, social media and science domains, we split the whole dataset into training, validation and test sets by ourselves since the original works do not specify how to split these datasets or the published datasets do not contain the split training, validation and test sets. Since the dataset sizes are limited for science, movie review and dialog domains, the maximum training samples for these domains are…”

A Generative Approach to Titling and Clustering Wikipedia Sections

“…In our Wikipedia section heading generation task, the prevalence of generic headings makes the task more abstractive than datasets like Gigaword (Rush et al, 2015), or even other short-text generation tasks, like email subject prediction (Zhang and Tetreault, 2019), which makes it a useful dataset for analyzing model performance. It is also extrinsically useful -most automated methods for improving Wikipedia focus on creating new content, such as through multi-document summarization or generating text from structured data (Lebret et al, 2016).…”

Proceedings of the Fourth Workshop on Neural Generation and Translation

“…In our Wikipedia section heading generation task, the prevalence of generic headings makes the task more abstractive than datasets like Gigaword (Rush et al, 2015), or even other short-text generation tasks, like email subject prediction (Zhang and Tetreault, 2019), which makes it a useful dataset for analyzing model performance. It is also extrinsically useful -most automated methods for improving Wikipedia focus on creating new content, such as through multi-document summarization or generating text from structured data (Lebret et al, 2016).…”