Bastian Seifert

ADDRESS Universitätstrasse 6, 8092 Zürich, Switzerland

Bastian Seifert is a postdoc of Markus Püschel. Before joining ETH he did his PhD under the supervision of Knut Hüper at the department of mathematics of the Julius-Maximilians-Universität Würzburg, while being a research assistant at the Center for Signal Analysis of Complex Systems of the University of Applied Sciences Ansbach. His main research interest is in algebraic signal processing theory. He did some research in the area of dimensionality reduction of epileptic EEG data, as well.

Co-advised theses at ETH

  • Fast Möbius and Zeta Transforms
    Tommaso Pegolotti, Master’s thesis 2021 (with Markus Püschel)

Publications related to algebraic signal processing


Porting Signal Processing from Undirected to Directed Graphs: Case Study Signal Denoising with Unrolling Networks
Vedran Mihal, Bastian Seifert, Markus Püschel
to appear in Proc. 30th European Signal Processing Conference (EUSIPCO)

Learning Fourier-Sparse Functions on DAGs
Bastian Seifert, Chris Wendler, Markus Püschel
ICLR 2022 Workshop on the Elements of Reasoning: Objects, Structure and Causality [poster]


Discrete Signal Processing on Meet/Join Lattices
Markus Püschel, Bastian Seifert, Chris Wendler
IEEE Transactions on Signal Processing, Vol. 69, pp. 3571-3584, 2021 [code]

Ranking Lattice
Wiener Filter on Meet/Join Lattices
Bastian Seifert, Chris Wendler, Markus Püschel
Proc. International Conference on Acoustics, Speech, and Signal Processing (ICASSP), pp. 5355-5359, 2021 [poster]
Generalized coverage function
Learning Set Functions that are Sparse in Non-Orthogonal Fourier Bases
Chris Wendler, Andisheh Amrollahi, Bastian Seifert, Andreas Krause, Markus Püschel
Proc. AAAI Conference on Artificial Intelligence, pp. 10283-10292, 2021

Digraph boundary conditions
Digraph Signal Processing with Generalized Boundary Conditions
Bastian Seifert and Markus Püschel
IEEE Transactions on Signal Processing, Vol. 69, pp. 1422-1437, 2021 [code] [slides+poster for ICASSP 22]


Undirected C2
The discrete cosine transform on triangles
Bastian Seifert and Knut Hüper
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Decomposed C2 Lattice
FFT and orthogonal discrete transform on weight lattices of semi-simple Lie groups
Bastian Seifert
submitted for publication


Fast cosine transform for FCC lattices
Bastian Seifert, Knut Hüper and Christian Uhl
13th APCA International Conference on Automatic Control and Soft Computing (CONTROLO)

Publications related to dimensionality reduction of epileptic EEG data


Subspace Detection and Blind Source Separation of Multivariate Signals by Dynamical Component Analysis (DyCA)
Christian Uhl, Moritz Kern, Monika Warmuth and Bastian Seifert
IEEE Open Journal of Signal Processing


Dynamical Component Analysis (DyCA) and its application on epileptic EEG
Katharina Korn, Bastian Seifert and Christian Uhl
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)


Dynamical Component Analysis (DyCA): dimensionality reduction for high-dimensional deterministic time-series
Bastian Seifert, Katharina Korn, Steffen Hartmann and Christian Uhl
IEEE 28th International Workshop on Machine Learning for Signal Processing (MLSP)

Analytic quantification of Shilnikov chaos in epileptic EEG data
Bastian Seifert, Dennis Adamski and Christian Uhl
Front. Appl. Math. Stat.,

See also the Google Scholar site.


PhD in Mathematics, Julius-Maximilians-Universität Würzburg, Germany
advisor Knut Hüper
Thesis: Multivariate Chebyshev polynomials and FFT-like algorithms
June 2019 (thesis handed in) / June 2020 (thesis defense)
Master (M.Sc.) in Mathematics, Julius-Maximilians Universität Würzburg, Germany
advisor Stefan Waldmann
August 2015
Bachelor (B.Sc.) in Mathematics, Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany
August 2013


Otto-Volk-Urkunde of the Otto-Volk-Stiftung for top grade Master degree 2016