Job Talk - Special Faculty

  • Postdoctoral Associate
  • Language Technologies Institute
  • Carnegie Mellon University

Domain Problems, Hypotheses, Creative Solutions: Teaching Data Science Through Applied Research

Analytical data science happens at the intersection of domain problems, data, and statistical models. It requires a combination of practical problem-solving skills and the ability to conduct rigorous hypothesis-driven experiments in order to fully utilize its potential. Working on real-world problems often entails additional obstacles such as sparse or noisy data, steep domain learning curves, or less well-defined tasks. This talk has three objectives. First, I illustrate how I have tackled these challenges in my research on artificial intelligence and law, specifically in computational models of legal argument and legal information retrieval. Second, I show how teaching data science at the Master’s level benefits from having students engage with hard research problems and domain experts. Third, I explain how applied research and data science education can synergize in the context of my work in the LTI’s Master of Computational Data Science (MCDS) program. On this basis, I outline a vision and plan for future developments of this degree program.

Matthias Grabmair has been a postdoctoral associate at CMU LTI working with Prof. Eric Nyberg since the Fall of 2015. His research focuses on various aspects of Artificial Intelligence & Law, question answering, and dialogue systems. He co-teaches seminar courses and mentors Capstone projects for the Analytics major of LTI’s MCDS program. Matthias obtained a diploma in law from the University of Augsburg, Germany, as well as a Master of Laws (LLM) and PhD in Intelligent Systems from the University of Pittsburgh.

Systems Scientist

For More Information, Please Contact: