Research

Below is an overview of the technical areas currently being researched by the Minds, Machines, and Society group:

Natural Language Processing & Machine Learning

Our work on natural language processing and machine learning includes:

  • Developing interpretability methods for Large Language Models
  • Guided natural language generation using Large Language Models
  • Detecting and mitigating antisocial tendencies of Large Language Models, such as bias and lack of alignment with human moral values
  • Using Transformer-based models
  • Understanding, modeling, and transferring style in text
  • Exploring prompt-based and few-shot learning

Social Media Analytics, Computational Social Science, and Graphs

Our work in social media analytics, computational social science, and graphs includes:

  • Extracting information from noisy user-generated text
  • Modeling user behavior (e.g., behavioral modeling, demographic/psychometric classification, etc.)
  • Modeling malicious behavior (e.g., influence campaigns, propaganda, etc.)
  • Modeling information diffusion in networks
  • Using representation learning on graphs