Date of Award
Master of Science (MS)
Events and relations between them are an important aspect of understanding natural language. This task can be broken into two, extracting events and extracting relations between them. The first task is well studied over various domains. In my thesis, I focus on the second task i.e. the extraction of relations between events. This task has been traditionally modeled as a pairwise classification task with handcrafted features. I suggest a neural network model, with ideas borrowed from models used to solve coreference resolution tasks, to solve this problem. I also suggest a structured prediction model for solving this task.
Singh, Siddharth, "Extracting Semantic Relations Between Events" (2017). Computer Science Graduate Theses & Dissertations. 157.