The Deep Learning Corner

In Summer 2018 several institutes of the Faculty of Informatics, TU Wien, together with started to jointly develop a platform specifically dedicated to research in Deep Learning. Since then the activity has grown into a substantial research project currently dealing with diverse subjects, such as for example,

  • DL of point clouds (Wimmer group, Erler P, Celarek A, Gilmutdinov I, Fraiss S)
  • Anime DL (Wimmer group, Cardoso J, Hanko D, Kugler F, Thurner Y)
  • DL in drug design (Kaufmann group, Kan P)
  • Agricultural time series analysis with DL (Rauber group, Neubauer T,  Raubitzek S, Peesapati K)
  • Federated/adversarial ML/DL (Rauber group, Mayer R, Jankovic A, Pustozerova A, Llugiqi M)
  • DL in computer vision (Sablatnig goup, Zambanini S, Keglevic M)
  • DL in e-commerce and NLP (Hanbury group, Hofstaetter S, Althammer S)
  • Research Unit ML/DL (Gaertner group, Holzmueller M, Thiessen M)
  • DL in SoC (Jantsch group, Wuschnig M)
  • DL in Intelligent Semantic Multimedia Data Analysis (Eidenberger group, Rybnikova T, Othman A)
  • DL for predicting water quality/quantity of karstic springs (Blaschke group, Derx J, Poelz A)

4 servers with 4 x GPU (NVIDIA Turing rtx2080ti) each
1 server with 1 x GPU (NVIDIA Volta V100)
accessible from the login node of the datalab.


The DL group has developed an exceptional high level of self-organization. New users will rapidly get introduced into established work policies and receive first level support from research fellows actively involved in DL research on a daily basis. This not only includes best practice advice on optimal usage of common DL frameworks (pytorch, tensorflow, keras, caffee, theano...) but also general assistance in mastering DL compute servers and adopting the required skill set to make efficient use of the available resources. In regular intervals a seminar series is run featuring current topics in DL as outlined by invited guest lecturers in the field.

Early success stories

Philipp Erler, Paul Guerrero, Stefan Ohrhallinger, Michael Wimmer, Niloy Mitra; Points2Surf: Learning Implicit Surfaces from Point Clouds in Computer Vision -- ECCV 2020, pages 108-124. October 2020.

Sebastian Hofstaetter, Hamed Zamani, Bhaskar Mitra, Nick Craswell, Allan Hanbury;
Local Self-Attention over Long Text for Efficient Document Retrieval in Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval -- SIGIR '20, pages 2021–2024. July 2020.


For informal discussions interested persons should best get into contact with any of the above research groups . If accounts are requested please write email to providing a valid TU email address and ideally an affiliation with any of the above groups.

  • No labels