Siinviidatu võib esmapilgus tunduda punase autona, st justkui ei kuulu siia. Ja kes nii soovib arvata, siis mis saab minul selle vastu olla. Kui igapäevasele minuni jõudvale e-posti liiklusele, siis üsna sageli jääb mulje lihtsalt ajaraiskamisest või -sisustamisest. Ja see on OK, kui osalised seda soovivad. Kui aga keegi soovib e-kirjad enda kasuks tööle panna, siis kulub veidi süsteemsem abi ära. Siinviidatu on vabalevis olev tekst, mis küll ei paku valmis äppi, kuid kust leiab hulgaliselt viiteid sellele, missugust tarkvara võiks otsima ja kasutama hakata.

Kontekstiks:

Automatic email to-do generation is the task of summarizing to-do items from given emails (Mukherjee et al., 2020) to help people overview overwhelming numbers of emails they receive every day (Radicati and Hoang, 2011) and schedule their daily work.

Autorid seavad sihte:

To fill in these gaps, in this work, we propose a learning to highlight and summarize model, where we learn the important sentence identification module and to-do summarization module concurrently in an end-to-end manner to focus on the most salient actions, as well as to incorporate structured action representations to generate more faithful todos.

Vihje:

To generate faithful to-do items that correlate actions with correct users from emails 1 , we propose to encode both text-level and action-level information as guidance signals.

Lugemishuvi suurendamiseks:

In this work, we propose a simple yet effective learning to highlight and summarize framework (LHS) to learn to identify salient text and actions from both email text and the constructed action graph, and generate faithful to-do items jointly.

Zhang, K., Chen, J., & Yang, D. Focus on the Action: Learning to Highlight and Summarize Jointly for Email To-Do Items Summarization.