Anastasiia Sedova

Anastasiia Sedova

NLP PhD Student in the Research Group Data Mining and Machine Learning, University of Vienna

About

I am a fourth-year PhD student in the Research Group Data Mining and Machine Learning at the University of Vienna, where I am advised by Ben Roth, working on machine learning and natural language processing. Previously, I completed my M.Sc. degrees in Computational Linguistics and Computer Science in the Center for Information and Language Processing at LMU Munich with a full scholarship from the German Academic Exchange Service. My research interests are learning with noisy and ambiguous data, data-centric methods for language models, model explainability and variability.

News

May 2024: I will be at ICLR’24. Let’s catch up!

April 2024: I gave an invited tutorial at the Department of Statistics at LMU Munich.

January 2024: I will be visiting Hinrich Schütze’s Lab at LMU Munich for one month in February 2024!

December 2023: I will be in person at EMNLP 2023 in Singapore! Also, I will be presenting our ULF paper at the poster session Sunday morning, 9am, East Foyer.

October 2023 : Our paper “ULF: Unsupervised Labeling Function Correction using Cross-Validation for Weak Supervision” was accepted to EMNLP 2023! Check it out here.

September 2023 : Invited tutorial on weak supervision at PhD Seminar “Machine Learning with Limited Data” at Aalborg University in Copenhagen. Check out the materials.

September 2023: I gave an oral presentation at ECML PKDD 2023 in Turin presenting our paper “Learning with Noisy Labels by Adaptive Gradient-Based Outlier Removal”.

June 2023: Our paper “Learning with Noisy Labels by Adaptive Gradient-Based Outlier Removal” was accepted to ECML PKDD 2023! (oral & poster) Check out paper and poster.

May 2023: Our paper “ACTC: Active Threshold Calibration for Cold-Start Knowledge Graph Completion” was accepted to ACL 2023! (poster) Check out paper and poster.

January 2023: Invited talk at the Technical University of Vienna on the potential application and future perspectives of weak supervision.

January 2022: I attended the ALPS2022 winter school on natural language processing organized by the University of Grenoble Alpes and Naver Labs Europe.

November 2021: Knodle framework presentation at the 3d Workshop of European Language Resource Coordination (ELRC).

August 2021: I presented our paper “Knodle: Modular Weakly Supervised Learning with PyTorch” online at ACL 2021!

August 2021: I attended the OxML’21 summer school organized by AI for Global Goals in partnership with CIFAR and the University of Oxford. The program included talks by Yoshua Bengio, Sebastian Ruder, Michael Bronstein and others. Acceptance rate: ~5%.

July 2021: This summer I’m taking part in Google Summer of Code. I will join DeepPavlov team to work on relation extraction in their end-to-end chatbot. Upd: check out my Medium post!

July 2021: Invited talk at Munich NLP Meetup introducing our freshly-baked Knodle!

June 2021: Our paper “Knodle: Modular Weakly Supervised Learning with PyTorch” was accepted to the Workshop on Representation Learning for NLP (RepL4NLP) hosted at ACL 2021! (poster) Check it out here.

May 2021: Invited talk at the Bosch NLP and Semantic Reasoning Lab presenting Knodle.

April 2021: Knodle first release!! Check out the source code.

I am also happy to serve as a reviewer at the international conferences, such as ACL, EMNLP, ECML, and others.

Selected Publications

Thesis supervision

  • Benedikt Köhler (11/2022-) Political Leaning Prediction of News Paper Articles, MSc thesis.
  • Sarah Sulollari (10/2022- ) - Analysing political developments by predicting party affiliation from Austrian parliament speeches, MSc thesis.
  • Lena Zellinger (01/2022-08/2022) - Gradient Matching for Learning with Noisy Data, MSc thesis.