Natural Language Processing for Society

Most societal applications involve multiple stakeholders. An important characteristic of such applications is the plurality of opinions on all aspects of the application from requirements to the desired outcomes. Language is the primary means in which the stakeholders express their opinions. Thus, language technology has an important role in facilitating constructive dialog and deliberation in such applications.

NLP for Society course teaches concepts and methods necessary for building dialog and deliberation-support systems that can facilitate societal exchange. The course facilitates learning via three avenues.

Lectures: The lectures in the course aim to introduce important societal challenges that language technology can help address, state-of-the art methods for addressing these challenges, key results, and open challenges. Examples of challenges covered include: (1) opinion analysis, (2) misinformation detection, (3) toxicity detection, and (4) value alignment.

Group Project: Complementing the lectures, students are expected to work on a group project, where they develop and/or evaluate an NLP method to address a societal challenge of their choice.

Literature Review: In addition to the group project, where the focus is on methods and evaluation, students are expected to work solo on a literature review of an application domain of societal relevance. Examples of relevant application domains include health, sustainability, democracy, education, mobility, and so on.

Together, these three avenues help students have a broad understanding of several societal challenges, dive deep into one of the challenges, and also have understanding of an application domain.