bwsmall.jpeg Email: lukas.pfahler tu-dortmund.de
Phone: 0231/755-8229
Fax: 0231/755-5105
Room-No.: OH12 R4.017

Consultation hour:
Thursdays, 14-16h. Please send me an e-mail ahead of time.
During the semester break (solely by arrangement)


Lukas Pfahler has studied computer science at the TU Dortmund University, where he received his Master of Science degree with distinction. Since 2016 he is a research associate at the chair for artificial intelligence at the department of computer science.

Research Interests

  • Statistical Learning Theory
  • Theoretical Foundations of Deep Network Learning
  • Embedding Learning



  • Visually Inspecting Singular Values in Deep Networks (Abstract)
  • Generalization in Deep Networks -- A very short introduction (Abstract)
  • What do you do with 5 million posts?: Versuche zum distant reading religiöser Online-Foren (Abstract)

Social Media

bitbucket.org | github.com | linkedin.com




Pfahler/Morik/2018a Pfahler, Lukas and Morik, Katharina. Nystroem-SGD: Rapidly Learning Kernel-Classifiers with Conditioned Stochastic Gradient Descent. In Machine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2018, Dublin, Ireland, 2018.
Pfahler/etal/2017a Pfahler, Lukas and Morik, Katharina and Elwert, Frederik and Tabti, Samira and Krech, Volkhard. Learning Low-Rank Document Embeddings with Weighted Nuclear Norm Regularization. In Proceedings of the 4th IEEE International Conference on Data Science and Advanced Analytics, 2017.
Buschjaeger/etal/2015a Buschjäger, Sebastian and Pfahler, Lukas and Morik, Katharina. Discovering Subtle Word Relation in Large German Corpora. In Proceedings of the 3rd Workshop on the Challenges in the Management of Large Corpora, 2015.
Morik/etal/2015a Morik, Katharina and Jung, Alexander and Weckwerth, Jan and Rötner, Stefan and Hess, Sibylle and Buschjäger, Sebastian and Pfahler, Lukas. Untersuchungen zur Analyse von deutschsprachigen Textdaten. No. 2, Technische Universität Dortmund, 2015.
Pfahler/2015a Pfahler, Lukas. Explicit and Implicit Feature Maps for Structured Output Prediction. TU Dortmund, 2015.
Pfahler/2013a Pfahler, Lukas. Effizienteres k-means Clustering von Zeitreihen mit Dynamic Time Warping durch kaskadiertes Berechnen von unteren Schranken. 2013.


Supervised Theses