Teaching 2020-2021:

  • CA377 Programming Fundamentals Project (Semester 1)
  • CA278 Programming Fundamentals III (Semester 2)
  • CA4023 Natural Language Technologies (Semester 2)

Final Year Project and M.Sc. Dissertation Supervision 2020-2021:

I am interested in supervising projects in the area of Natural Language Processing. Here are some ideas:
  • Story Generation: can neural nets generate believable stories?
  • Web app development: building an interactive website to showcase the research on Irish Language Technologies done by researchers in DCU's School of Computing (suitable for CA students who like front-end development)
  • Aspect-based sentiment analysis of product reviews
  • Explainable AI: how to explain and visualise the decisions of a deep neural net?
  • Machine Reading Comprehension: can a neural net understand and answer questions about a text?

Publications:

For an up-to-date list of my publications, please see my Google Scholar profile

PhD Students:

Current Students
  • Daria Dzendzik (Automatic Reading Comprehension)
  • Henry Elder (Neural Language Generation)
  • James Barry (Univeral Dependency Parsing)
  • Abigail Walsh (Multiword Expressions in Irish)
  • Lauren Cassidy (POS Tagging and Parsing of Irish Twitter)
  • Alan Cowap (Emotions in Automatic Text Generation)
  • Chenyang Lyu (Pretrained language models)
Former Students
  • Piyush Arora (Learning Through Search)
  • Utsab Barman (Code-Switching in Social Media Communication)
  • Dasha Bogdanova (Forum Question-Answering)
  • Rasoul Kaljahi (Syntax and Semantics for Machine Translation Quality Estimation)
  • Teresa Lynn (Treebanking and Dependency Parsing for the Irish Language)
  • Joachim Wagner (Detecting Grammatical Errors with Probabilistic Parsers)

Resources:

  • Foreebank: Hand-annotated constituency trees with grammatical error corrections for 1,000 English and 1,000 French Norton Forum sentences (with Rasoul Kaljahi and Johann Roturier). Paper
  • SymForum: An English/French Norton Forum dataset for machine translation quality estimation (with Rasoul Kaljahi and Johann Roturier).  Paper
  • Universal Dependencies for Irish: 1,000 Irish sentences annotated in the universal dependencies scheme (with Teresa Lynn).
  • GenERRate: A tool for automatically inserting a grammatical error into a sentence.  Paper. GitHub Repository (thanks to Nicholas Dronen)
  • BNC Test Set: PTB-style hand-annotated phrase structure trees for 1,000 BNC sentences.  Paper