Pauli Miettinen
Professor of Data Science
Email:
pauli.miettinen@uef.fi
Phone: +358 2 9445 3410
Fax: +358 2 9445 7318
Who am I?
I am a professor of data science at UEF's School of Computing, at Kuopio campus. Formerly I was a senior researcher at Max Planck Institute for Informatics in Saarbrücken, Germany. I am also a docent (adjunct professor) at University of Helsinki.
I study algorithmic data analysis, and I am interested to both theory and applications thereof. I lead the Algorithmic Data Analysis research group at UEF. My main research has been in the fields of Boolean and (sub-) tropical matrix and tensor factorizations, redescription mining, and social network analysis. You can find my publications in my Publications page.
What's new
- Our paper Hashing for Fast Pattern Set Selection will appear in ECML PKDD 2025
- Jonne Klockars joins the group as the first hire to PAPADAM-project
- Research Council Finland funds the PAPADAM-project, co-PIs me and Nikolaj Tatti from University of Helsinki
- Interesting abstracts from interesting Dagstuhl seminar here.
- Two worlds collide: our work using NMF to describe redescriptions has appeared in IEEE TKDE.
- Charmin Asirimath joins the group as a part of the Finnish Doctoral Program Network in Artificial Inteligence (AI-DOC). Welcome!
- Ambre Ayats joins the group as a YUFE postdoc. Welcome!
- How to mine redescriptions faster than before? Check out the new paper to appear at ECML PKDD!
- Ever wondered how a Boolean matrix factorization looks like? Wonder no more!
- How to mine differentially private redescriptions with a greedy algorithm? What about differentially private tree-based redescriptions? Check out our new papers!
Recent publications
-
Hashing for Fast Pattern Set Selection.
Proc. ECML PKDD,
,
[manuscript | tech. rep. | source code] -
Hashing for Fast Pattern Set Selection.
arXiv:2507.08745 [cs.DB]
.
10.48550/arXiv.2507.08745
[pdf (arXiv) | source code] -
Hyperbolic Communities: Modelling Communities Beyond Cliques.
Statistical and Probabilistic Methods in Algorithmic Data Analysis (Dagstuhl Seminar 24391)
14(9),
,
136.
10.4230/DagRep.14.9.127
[pdf (Schloss Dagstuhl)] -
Finding Rule-Interpretable Non-Negative Data Representation.
IEEE Transactions on Knowledge Discovery and Data Mining
37(5),
,
2538–2549.
10.1109/TKDE.2025.3538327
[ pdf (IEEE) | tech. rep. | source code] -
Fast Redescription Mining Using Locality-Sensitive Hashing.
Proc. ECML PKDD,
,
124–142.
10.1007/978-3-031-70368-3_8
[manuscript | tech. rep. | source code] -
Fast Redescription Mining Using Locality-Sensitive Hashing.
arXiv:2406.04148 [cs.LG]
.
10.48550/arXiv.2406.04148
[pdf (arXiv) | source code] -
Visualizing Overlapping Biclusterings and Boolean Matrix Factorizations.
Proc. ECML PKDD,
,
743–758.
10.1007/978-3-031-43412-9_44
[manuscript | pdf (Springer) | tech. rep. | source code] -
Visualizing Overlapping Biclusterings and Boolean Matrix Factorizations.
arXiv:2307.07396 [cs.LG]
.
10.48550/arXiv.2307.07396
[pdf (arXiv)] -
Differentially Private Tree-Based Redescription Mining.
Data Mining and Knowledge Discovery
37(4)
,
1548–1590.
10.1007/s10618-023-00934-8
[pdf (Springer) | tech. rep. | source code] -
Serenade: An Approach for Differentially Private Greedy Redescription Mining.
Proc. 20th Anniversary Workshop on Knowledge Discovery in Inductive Databases,
,
31–46.
[manuscript | pdf (CEUR-WS) | source code] -
Differentially Private Tree-Based Redescription Mining.
arXiv:2212.06630 [cs.DB]
.
10.48550/arXiv.2212.06630
[pdf (arXiv)] -
Finding Rule-Interpretable Non-Negative Data Representation.
arXiv:2206.01483 [cs.LG]
.
10.48550/arXiv.2206.01483
[pdf (arXiv)]