Alexander Jung
Assistant Professor of Computer Science, Aalto University
Contact information
My research group
Publications
Teaching
Short CV
My Top Achievements
"There is Nothing More Practical Than A Good Theory" (Kurt Lewin)
Compressed Sensing over Complex Networks for (SemiSupervised) Learning from Big Data (CSLeBiD)
My current research revolves around applying the theory and algorithmics of compressed sensing to massive datasets with intrinsic network structure, i.e., to big data over networks. In particular, I study relations between concepts of the theory of complex networks (e.g., clustering, percolation) and the theory of sparse graph signals defined over such complex networks. The combination of compressed sensing with modern clustering algorithms is expected to be game changing for (semisupervised) learning from big data over networks in a similar manner as compressed sensing was for digital signal processing.
Upcoming
Lecture 6 of our Convex Optimization course next Wed. (29.3.2017) in lecture hall M232 (M1) at
Otakaari 1. Convex Optimization for Graphical Models
Selected Publications

M. Hinkka, T. Lehto, K. Heljanko and A. Jung. Structural Feature Selection from Event Logs. submitted , Mar. 2017.

A. Jung, A. Heimowitz and Y. C. Eldar. The Network Nullspace Property for
Compressed Sensing over Networks. submitted , Feb. 2017. Preprint

N. Tran Quang and A. Jung. On the Sample Complexity of Graphical Model Selection for NonStationary Processes. submitted , Jan. 2017. Preprint

A. Jung. Sparse Label Propagation. submitted to JMLR, Dec. 2016. Preprint: arXiv:1612.01414 [CS.LG]

A. Jung, A. Hero III, A. Mara and S. Aridhi. Scalable SemiSupervised Learning over Networks using Nonsmooth Convex Optimization. submitted , Oct. 2016. Preprint: arXiv:1611.00714 [CS.LG]
 A. Jung, Y. C. Eldar, and N. Goertz. On the Minimax Risk of Dictionary Learning. IEEE Trans. Inf. Theory, vol. 62, no. 3, March 2016. Preprint: arXiv:1507.05498 [stat.ML]

A. Jung. Learning the Conditional Independence Structure of Stationary Time Series: A Multitask Learning Approach. IEEE Trans. Sig. Proc., to appear.
Preprint: arXiv:1404.1361 [stat.ML]

A. Jung, G. Hannak, and N. Goertz. Graphical LASSO based Model Selection for Time Series. IEEE Sig. Proc. Letters, October 2015. Preprint: arXiv:1410.1184 [stat.ML]
Selected Talks

Sparse Label Propagation
in group seminar of Prof. Thomas Pock at TU Graz, April 2017.

Compressed Sensing for Learning from Big Data over Networks
at International Institute for Applied Systems Analysis (IIASA) Austria, March 2017.

Compressed Sensing for Learning from Big Data over Networks
at Johannes Kepler University (JKU) Linz, Feb. 2017.

Compressed Sensing for Learning from Big Data over Networks .
within the
Large Structures Seminar of Aalto University, Feb. 2017.

Compressed Sensing for SemiSupervised Learning from Big Data over Networks.
within the
Machine Learning Coffee Seminars of Aalto University and the University of Helsinki, Feb. 2017.

Compressed Sensing for Big Data over Networks. in group seminar of Prof. Sjoerd Dirksen and Prof. Holger Rauhut at RWTH Aachen, Jan. 2017.
 Graphical Model Selection for Big Data over Networks. within the Stochastic Sauna at Aalto University,
Dec. 2016.
 Graph Signal Recovery using Convex Optimization. within the
Helsinki Algorithms Seminar of Aalto University and the University of Helsinki, May 2016.
 An InformationTheoretic Approach to Dictionary Learning. TU Berlin, June 2015.
 On the SampleComplexity of Dictionary Learning. within the Mathematics Colloquium of the University of Innsbruck, June 2015.
 Performance Limits of Dictionary Learning for Sparse Coding: An InformationTheoretic Approach. University of Michigan, June 2014.
 Compressive Nonparametric Graphical Model Selection for Time Series: A Multitask Learning Approach. The University of Edinburgh, Oct. 2013.
first.last at aalto.fi