JSP3

Semester project 3 for Linguistics

Goals

• analyze some text
• learn techniques for analyzing text
• learn some software engineering skills
• learn git, markdown, use LaTeX

2026 task: TBD

• how many tasks? 4

• how many teams? 4

• Introduce yourselves

  • name
  • languages spoken (natural/programming)
  • interests (relevant to the course)
  • one fun fact (not relevant to the course)

What should the task be?

• Body-focused repetitive behaviour (BFRB) Sara

  • look at Doctor-patient interaction
  • look for examples of BFRB in literature/media
  • TASK:
    • READ A Qualitative Study Exploring International Experiences of Seeking Treatment for Adults With Trichotillomania: A Story of Frustration and Unmet Need
    • READ Self-control and Body-focused Repetitive Behaviors
    • look at Books about Trichotillomania and Dermatillomania could we analyze these? Are there similar books in Czech?

• How women are represented online. Anna

  • For example in news articles and user comments and maybe comparing different time periods to see how the language changed over time.
  • I would also like to compare how men and women talk about each other online.
  • TASK: find some literature, refine the problem

• Distribution of lexicalized metaphor Erika

  • combine sense-tagged text with ChainNet
  • analyze distribution
  • fill in what is missing
  • TASK: Read these papers:
    • ChainNet: Structured Metaphor and Metonymy in WordNet
    • Teaching Through Tagging — Interactive Lexical Semantics

• Improve Czech wordnet Zuzana

  • fill in Czech specific patterns
    • adverbs [of languages]
    • diminutives
    • aspect variants
    • gender variants (role nouns)
  • generate definitions for derived entries
  • get examples from corpora
  • derivational links with affixes?
  • READ Papers:
    • Czech Wordnet 1.9 (note we have a private copy of Czech Wordnet 2.1 from Adam Rambousek).
    • Derivational Relations in Czech WordNet
    • Overview and Future of Czech Wordnet (note the open release did not happen)
    • Lexico-Semantic Annotation of the Prague Dependency Treebank
    • https://deb.fi.muni.cz:8005/debdict/
    • https://cygnet.maudslay.eu/
    • https://pypi.org/project/wn/

• LLM text vs Human text

  • what is different (test for Czech?)
  • I have data for English
  • TASK: Read these papers:
    • Identification of AI-generated texts using quantitative markers
    • More Aligned, Less Diverse? Analyzing the Grammar and Lexicon of Two Generations of LLMs

Project goals

• get something useful for each task
• combine to make a best-of-breed
• write, submit and publish a paper
• release at least one automatically tagged, aligned corpus

Approach

• use github to coordinate
  • GitHub quickstart
  • git an Introduction

Weekly meeting

• Erika and Anna Monday

• Sara and Zuzana Wednesday (3pm)

• 15-30 minutes progress

• longer discussion of issues as necessary

Next tasks

• make github account
  • send me accountname
  • I will add to github
• Background reading, refine problems