Computer Science

Machine Learning and Data Mining

Andreas Maunz received his masters in computer science from Albert-Ludwigs University, Freiburg. He is currently a PhD student at the chair of Prof. Stefan Kramer. His main interests are machine learning and data mining in large amounts of data using (sub)symbolic representation, especially mining and modeling of discrete data with pattern recognition techniques, including frequency- and generality-based as well as statistical filters.

His core competences lie in 2D graph based representations providing a good trade-off between expressiveness and computational complexity. This combines methods such as graph mining, supervised feature selection, dimensionality reduction, as well as activity-specific similarity indices. The current Lazar prediction system offers automatic applicability domain estimation and full details for every single prediction to expert users from the biochemical domain in a user friendly web interface.

Reliable and user friendly computer-based models and alerts in drug discovery and risk management, particularly for early stages of development, are high priority for both regulators and the pharmaceutical industry, especially with regard to endpoints which are not fully elucidated so far.
Interestingly, this constitutes a challenging field for computer scientists due to large amounts of structured data. There are many possible representations for chemicals and their properties that lend themselves to data mining tasks. Creating an inductive database therefore implies solutions to sophisticated problems such as graph mining, pattern recognition, and multivariate modeling with dimensionality reduction, also relevant to other areas of applied computer science.

Andreas Maunz is specifically concerned with methods to find structural alerts, i.e. fragments of chemical structures associated with biological activity, in a fully automated fashion. The objective is to establish an inductive database of compounds and activity values that users can query for fragments relevant to the endpoint of interest, providing, among others, frequency and statistical constraints. For this purpose efficient subgraph enumeration including pruning techniques for statistical and frequency constraints are employed.

Introduction: Click here to see a 70s example of machine learning in action: The artificial neural network nettalk (1987) learns to pronounce a specific english sentence.
Click here to see a fascinating modern example of cybernetic engineering: The Golem project, which is concerned with the creation of evolutionary robots.

Software

LibBBRC for mining backbone refinement class representatives [API, Supporting Website]
LibLAST for latent structure pattern mining, [API, Supporting Website]
Ready-to-use appliance with both BBRC and LAST, based on VirtualBox.
Multi-View Clustering (Bickel und Scheffer, ICDM 2004) in R (also available from CRAN).

Publications

Items not available as download may be obtained upon request. See also my indexed entries on PubMed and DBLP.

Journals and Book Chapters:
- Maunz A, Gütlein M, Rautenberg M, Vorgrimmler D, Gebele D, Helma C (2013). “Lazar: A Modular Predictive Toxicology Framework” Frontiers in Pharmacology 4(38). [bib, (Open Access)]
- Suenderhauf, C, Hammann, F, Maunz, A, Helma, C, and Huwyler, J (2011).
  “Combinatorial QSAR Modeling of Human Intestinal Absorption”
  Molecular Pharmaceutics, 8(1):213-224. [bib]
- Hardy, B, Douglas, N, Helma, C, Rautenberg, M, Jeliazkova, N, Jeliazkov, V, Nikolova, I, Benigni, R, Tcheremenskaia, O, Kramer, S, Girschick, T, Buchwald, F, Wicker, J, Karwath, A, Gütlein, M, Maunz, A, Sarimveis, H, Melagraki, G, Afantitis, A, Sopasakis, P, Gallagher, D, Poroikov, V, Filimonov, D, Zakharov, A, Langunin, A, Gloriozova, T, Novikov, S, Skvortsova, N, Druzhilovsky, D, Chawla, S, Gosh, I, Ray, S, Patel, H, and Escher, S (2010).
  “Collaborative Development of Predictive Toxicology Applications”
  Journal of Cheminformatics, 2(7). [bib, pdf (Open Acces)]
- Maunz, A and Helma, C (2010).
  “Prediction of Toxic Effects of Pharmaceutical Agents”
  In: Pharmaceutical Data Mining: Approaches and Applications for Drug Discovery, ed. by Konstantin V. Balakin, Sean Ekins. Wiley, New York, NY, USA, chap. 5, pp. 145-176. (ISBN: 978-0-470-19608-3). [abstract, bib]
- Hammann, F, Gutmann, H, Jecklin, U, Maunz, A, Helma, C, and Drewe, J (2009).
  “Development of Decision Tree Models for Substrates, Inhibitors, and Inducers of p-Glycoprotein”
  Current Drug Metabolism, 10(4):339-346. [bib]
- Maunz, A and Helma, C (2008).
  “Prediction of Chemical Toxicity with Local Support Vector Regression and Activity-Specific Kernels”
  SAR and QSAR in Environmental Research, 19(5-6):413-431. [pdf, bib]

Conferences (full proceedings):

Maunz, A, Vorgrimmler, D, Helma C. “Out-of-Bag Discriminative Graph Mining” In: Proceedings of the 28th Annual ACM Symposium on Applied Computing Coimbra, Portugal March 18-22, 2013.
Maunz A, Helma C, and Kramer S (2011).
“Efficient Mining for Structurally Diverse Subgraph Patterns in Large Molecular Databases”
In: Machine Learning, vol. 83(2), pp. 193-218, Springer Netherlands. [pdf, bib]
Maunz A, Helma C, Cramer T, and Kramer S (2010).
“Latent Structure Pattern Mining”
In: Balcazar, Jose; Bonchi,Francesco; Gionis, Aristides; Sebag, Michele: ECML/PKDD 2010: Machine Learning and Knowledge Discovery in Databases, vol. 6322, pp. 353-368, Berlin / Heidelberg, Springer. [pdf, poster pdf, bib, Supporting Website]

Maunz A, Helma C, and Kramer S (2009).
“Large Scale Graph Mining using Backbone Refinement Classes”
In: KDD ’09: Proceedings of the 15th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 617-626, New York, NY, USA, ACM.
[pdf, bib, Supporting Website]

© ACM, 2009. This is the author’s version of the work. It is posted here by permission of ACM for your personal use. Not for redistribution. The definitive version was published in KDD ’09: Proceedings of the 15th ACM SIGKDD international conference on Knowledge discovery and data mining, pages 617-626, 2009, http://doi.acm.org/10.1145/1557019.1557089.

Video lecture: slides 1-3.exe, 4-6.exe, 7-10.exe, 11-15.exe, 16-19.exe, 20-23.exe or view the conference recordings.

Conferences (presentations):
- Maunz A: “Elaborate Graph Mining: Exploiting Structural Invariants and Latent Information in Graph Databases to Predict REACH-Relevant Endpoints”, presented at the OpenTox 2011 conference, Summer 2011, Technical University Munich (9-13 August 2011) [pdf]
- Maunz A, Helma C, Cramer T, and Kramer S. “Latent Structure Pattern Mining”, presented at the eCheminfo Community of Practice InterAction Meeting, Summer 2010, Oxford University, Oxford, UK (01-05 August 2010).
- Maunz A, Helma C: “New Lazar Developments”, presented at the eCheminfo Community of Practice InterAction Meeting, Autumn 2008, Bryn Mawr College, Philadelphia (13-17 October 2008).
  [ppt, pdf]
- Maunz A: “Instance-based Regression Models for
  Quantitative Biological Activities using Support Vector Machines and Multilinear Models”, presented at the Scarlet Workshop on in silico methods for carcinogenicity and mutagenicity, Milano, April 2008.
  [pdf]

Unpublished Material

Related to Publications:
- Maunz A: “A Quantitative Extension to the Lazar Algorithm for the Prediction of Chemical Properties.” [pdf diplomarbeit.pdf].
- Maunz A: “On the Number of Backbone Refinement Classes”, derivation and proof of a formula for counting the number of BBRCs in a perfect binary tree. Comparison to the total number of subtrees [pdf bbrc-no.pdf].
- Maunz A: “On the Co-Occurrence and Diversity of Backbone Refinement Classes”, euclidean embedding of BBRC features and instances in 2D as well as feature similarity alike to ORIGAMI approach [pdf bbrc-rep.pdf].
- Maunz A: “Support Vectors and the Margin in a Nutshell”,
  brief outline of support vector theory and the
  concept of margin. (see ch. 1,2 and 7 of
  Schölkopf and Smola, 2002) [pdf sv-margin.pdf].

Machine Learning and Data Mining

Software

Publications

Unpublished Material

Links

Andreas Maunz