Principal Investigator: Professor Ng Hwee Tou, SoC

We live in a world of information explosion. Multiple news agencies put out news reports around the clock. On social media such as Facebook and Twitter, continuous streams of posts contributed by Web users appear non-stop and in ever increasing quantity. Making sense of this massive amount of information presents a great challenge. An important category comprises natural language texts describing some emerging events of interest, such as terrorist attacks (hostage taking, bombing, gun firing), natural disasters (earthquakes, tornadoes), corporate merger and acquisition, etc. Such an event occurs over time and comprises a sequence of related subevents in chronological order. For example, the April 2022 New York City Subway attack comprises attack at the subway, arrest of the suspect, and his subsequent prosecution. In this project, given a stream of natural language news documents, the task of event chain summarization produces an abstractive summary that provides a concise description of the chain of subevents constituting the main event.