Mayumi Ohta

Mayumi Ohta

Mayumi Ohta, M.A.

I am a PhD student under the supervision of Prof. Dr. Stefan Riezler and member of the Statistical NLP Group.

Research Interests

  • Convex and Non-convex Optimization
  • Generalization and Transferability
  • Adversarial Machine Learning

Curriculum Vitae

Education

2019- PhD Student, Computational Linguistics
Heidelberg University, Germany
2017 Master of Arts, Computational Linguistics
Heidelberg University, Germany
2014 Bachelor of Arts, Computational Linguistics
Heidelberg University, Germany
2011 Research Student, Information Studies
University of Tokyo, Japan

Employment

05/2019 – Research Associate (full time)
Institute for Computational Linguistics, Heidelberg University, Germany
10/2019 – 12/2019 Research Intern
NEC Labs Europe, Heidelberg, Germany
Information Retrieval, Slot Filling
01/2018 – 04/2019 Research Scientist (full time)
Nuance Communications Inc., Aachen, Germany
Language modeling
11/2016 – 06/2017 Working Student
Leibniz ScienceCampus, Heidelberg, Germany [ LiMo ]
Semantic role linking
08/2013 – 10/2016 Working Student
IT department at the Centre for East Asian Studies, Heidelberg, Germany [ ZO ]
Web administration and application development
08/2014 – 02/2015 Software Developer (full time)
Publications Office at the Karl Jaspers Centre, Heidelberg, Germany [ HeiUp ]
Digital publishing
03/2013 Intern
Department of Scalability and Performance, SAP Inc., Walldorf, Germany
12/2007 – 03/2009 Working Student
Educational Planning Office at University of Tokyo, Tokyo, Japan
Video editorial, MOOC contents development
04/2007 – 03/2009 Trainee
Suzuki Audio GmbH, Tokyo, Japan
Video editorial, Broadcast operations

Teaching

2020/21 Winter Generalization in Deep Learning
2020 Summer Mathematical Foundations of Computational Linguistics
2017 Summer Interactive Neural Machine Translation (Tutorial for IUED Studients)

Publications

  1. Tsz Kin Lam, Mayumi Ohta, Shigehiko Schamoni and Stefan Riezler
    On-the-Fly Aligned Data Augmentation for Sequence-to-Sequence ASR
    Proceedings of the 22th Annual Conference of the International Speech Communication Association (INTERSPEECH), Brno, Czech Republic, 2021
    pdf | bib 
    @inproceedings{lam2021,
  author = {Lam, Tsz Kin and Ohta, Mayumi and Schamoni, Shigehiko and Riezler, Stefan},
  title = {On-the-Fly Aligned Data Augmentation for Sequence-to-Sequence ASR},
  journal = {Proceedings of the 22th Annual Conference of the International Speech Communication Association},
  journal-abbrev = {INTERSPEECH},
  year = {2021},
  city = {Brno},
  country = {Czech Republic},
  url = {https://arxiv.org/abs/2104.01393}
}
  2. Mayumi Ohta, Nathaniel Berger, Artem Sokolov and Stefan Riezler
    Sparse Perturbations for Improved Convergence in Stochastic Zeroth-Order Optimization
    The 6th International Conference on Machine Learning, Optimization, and Data Science (LOD), Siena, Italy, 2020
    pdf | bib 
    @article{ohta2020,
  author = {Ohta, Mayumi and Berger, Nathaniel and Sokolov, Artem and Riezler, Stefan},
  year = {2020},
  title = {Sparse Perturbations for Improved Convergence in Stochastic Zeroth-Order Optimization},
  journal = {The 6th International Conference on Machine Learning, Optimization, and Data Science},
  journal-abbrev = {LOD},
  city = {Siena, Italy},
  url = {https://arxiv.org/abs/2006.01759}
}
  3. Artem Sokolov, Julian Hitschler, Mayumi Ohta and Stefan Riezler
    Sparse Stochastic Zeroth-Order Optimization with an Application to Bandit Structured Prediction
    arXiv preprint arXiv:1806.04458, 2018 (Preprint)
    pdf | bib 
    @inproceedings{sokolov2018,
  author = {Sokolov, Artem and Hitschler, Julian and Ohta, Mayumi and Riezler, Stefan},
  title = {Sparse Stochastic Zeroth-Order Optimization with an Application to Bandit Structured Prediction},
  journal = {arXiv preprint arXiv:1806.04458},
  note = {Preprint},
  year = {2018},
  url = {https://arxiv.org/abs/1806.04458}
}
  4. Julian Hitschler, Laura Jehl, Sariya Karimova, Mayumi Ohta, Benjamin Körner and Stefan Riezler
    Otedama: Fast Rule-based Pre-Ordering for Machine Translation
    The Prague Bulletin of Mathematical Linguistics (PBML), 106, 159–168, 2016
    pdf | bib 
    @article{hitschler2016b,
  author = {Hitschler, Julian and Jehl, Laura and Karimova, Sariya and Ohta, Mayumi and K\"{o}rner, Benjamin and Riezler, Stefan},
  title = {Otedama: Fast Rule-based Pre-Ordering for Machine Translation},
  journal = {The Prague Bulletin of Mathematical Linguistics},
  journal-abbrev = {PBML},
  number = {106},
  pages = {159--168},
  year = {2016},
  url = {http://www.cl.uni-heidelberg.de/~riezler/publications/papers/MTM2016.pdf}
}