Katerina Margatina

Sheffield, Dec. 2019

Google Scholar

Hi👋! I am a third (final) year Ph.D. student at the Computer Science department at the University of Sheffield🇬🇧. I am working on natural language processing and machine learning, I am a member of the SheffieldNLP group, and my Ph.D. is funded by an Amazon Alexa Fellowship. My two amazing supervisors are Nikolaos Aletras (main) and Loïc Barrault. My research mostly focuses on active learning, low-resource learning, uncertainty estimation, robustness and language modeling – but I’m fascinated by other topics as well!

Currently, I live in beautiful Copenhagen🇩🇰, as I am doing a research visit at the CoAStaL group. I am excited to be working with Anders Søgaard and the rest of the team on learning from disagreement, privacy and fairness!

In 2019, I was working as a Machine Learning Engineer at the awesome Greek startup DeepSea Technologies. Prior, I obtained my diploma on Electrical and Computer Engineering (integrated BEng & MEng) at the National Technical University of Athens (NTUA)🇬🇷. In my thesis I worked on transfer learning and attention-based conditioning methods for natural language processing, under the supervision of Alexandros Potamianos.

In my free time, I enjoy yoga, movies and long walks (with iced coffee). I love tennis, skiing, hiking, and bouldering️ (even though I am not good at it)! I am generally looking forward to the post-Covid era, when we will be able to enjoy travelling again.🦋


Nov 5, 2021 I’m very lucky to be able to fly to Punta Cana for the EMNLP conference to present our papers.🌴🐚🌊🍍🥥 Can’t wait to meet old and new friends in-person after such a long time!
Aug 26, 2021 Super happy to have 2 papers accepted at EMNLP 2021 (main conf.)!🎉
Jul 1, 2021 Officially moved to Copenhagen, ready to start my PhD research visit at CoAStaL!🤓

selected publications

  1. EMNLP

    ✨ Oral ✨

    Active Learning by Acquiring Contrastive Examples
    Margatina, K., Vernikos, G., Barrault, L., and Aletras, N.
    In Proceedings of the Conference on Empirical Methods in Natural Language Processing 2021
  2. EMNLP
    Frustratingly Simple Alternatives to Masked Language Modeling
    In Proceedings of the Conference on Empirical Methods in Natural Language Processing 2021
  3. Bayesian Active Learning with Pretrained Language Models
    Margatina, K., Barrault, L., and Aletras, N.
    In ArXiv 2021