News is essential in a democracy, yet, recent financial strains have threatened news outlets throughout the world. Computational approaches aim to lower the cost and make high-quality journalism feasible. In this talk, we'll discuss several key questions in computational journalism: (1) identifying stories worth reporting, (2) recommending sources, (3) adding relevant background and context, and (4) responding to ongoing developments. We will examine these tasks through the lens of journalistic decision-making and delve into the connections between these tasks and advancements in generative language modeling. For example, understanding how information changes in breaking news cycles can improve abstention rates in QA tasks; similarly, modeling source requirements for stories can improve strategies for multi-document retrieval. Ultimately, as large language models (LLMs) usher us into a new era of information, computational journalism offers a pathway to both preserve and embed human values into the way we communicate information.
Alexander Spangher is formerly a writer and data scientist at the New York Times and is pursuing his PhD in computer science at the University of Southern California. He focuses on computational journalism and is advised by Jonathan May, Emilio Ferrara and Nanyun Peng. He has collaborated extensively with Stanford Big Local News as well as Bloomberg News. His research is broad and has pursued the following side directions: he has worked at Microsoft Research under the mentorship of Eric Horvitz to detect misinformation. He has collaborated with EleutherAI to build state-of-the-art symbolic music models. Finally, he has collaborated with the MIT Plasma Science and Fusion Center (PFSC) to model disruptions in nuclear fusion reactions. His work has received an Outstanding Paper Award at NAACL 2022 and he is fortunate to be supported by a multi-year Bloomberg PhD Fellowship.