Natural Language Processing

The Natural Language Processing (NLP) working group focuses on understanding human language with computers, and we develop methods for extracting knowledge from text with statistical methods. One of our main interests are deep learning methods, and we combine them with linguistic knowledge and knowledge about the world (which can be expressed in knowledge graphs).

Current members

Former members

Students: How to do a project or a thesis with us?

The NLP group offers practicals (Praktikum 1, Praktikum 2, Data Science Project, and others) and thesis topics (Bachelor, Master) for several study programs (Computer Science, Data Science, Digital Humanities, and others).

Please contact us as early as possible, if you think you might want to do a project or thesis in the area of Machine Learning for NLP or Knowledge Graphs. (The earlier, the better the topical fit. E.g., you could contact us even at the beginning of your Master studies about a potential thesis, without any commitment, of course.)

The best way to contact us is to join one of my (Benjamin Roth) virtual coffee sessions (link), Tuesdays, 2pm.

You can also contact us by email (benjamin.roth -at- univie.ac.at), please include the following information:

  • up-to-date transcript of records
  • Whether you
    • are generally interested in Machine Learning for NLP and Knowledge Graphs (no particular topic in mind)
    • or, have own ideas for a topic, or an interest in a specific problem in ML / NLP / KGs
    • or, are interested in one of the topics below (more details in Moodle). To see the topics on the Moodle page, you need to be logged in to Moodle with your u:account, and add yourself (click "Einschreiben") to the linked Moodle course.

Currently open topics

  • A Data Analysis of Existing NLP Datasets for Implicit Language Phenomena
  • Advanced Weakly Supervised Sequence Labeling
  • Applying Weak Supervision Methods for LLM Calibration
  • Data Augmentation for Language-Model Pretraining Datasets
  • Data Extraction Attacks for NLP Systems
  • Evaluate LLMs for Thematic Analysis on Students Interaction with GPT-4o
  • Fine-tuning Language Models using Temporal Constraints
  • Gradient Dimensionality Reduction for Instance-based Explainability of LLMs
  • Tracking Text Influence in Financial Documents
  • What Exactly is Private Information in Text?
  • What Makes an Effective Persona Prompt?