mypic

I am an associate professor at the Amsterdam University Medical Center and consultant at HORAIZON.
Email AMC: e.levin[at]amc.uva.nl
Email HORAIZON: evgeni.levin[at]horaizon.nl


My research interests span a range of topics in statistical machine learning and its applications, with a particular emphasis on intelligible and multi-view learning algorithms. At the Medical Center, I am working on cardiometabolic disorders (such as diabetes, coronary artery disease) with focus on the identification and evaluation of biomarkers via novel machine learning techniques tailored to the question at hand.

CODE

Domain Intelligible Model. We adapt sparse Generalized Additive Model (sparse GAM), to be applicable to the task of variable selection in high dimensional, microbiome (-omics) dataset. GAMs are less general in comparison to "fully" nonparametric models, but have a notable advantage of being readily interpretable and easier to estimate using a simple backfitting algorithm. Recently, standard additive models have been successfully applied in the biomedical domain, and they can be naturally extended to include various interactions among predictors.

Co-regularized sparse-group lasso. We introduce the co-regularized sparse-group lasso algorithm: a technique that allows the incorporation of auxiliary information into the learning task in terms of groups of predictors and the relationship between those groups. The proposed cost function requires related groups of predictors to provide similar contributions to the final response, and thus, guides the feature selection process using auxiliary information. Our algorithm is particularly suitable for a wide range of biological applications where good predictive performance is required and, in addition to that, it is also important to retrieve all relevant predictors so as to deepen the understanding of the underlying biological process.

Unsupervised multi-view feature selection via co-regularization. Existing unsupervised feature selection algorithms are designed to extract the most relevant subset of features that can facilitate clustering and interpretation of the obtained results. However, these techniques are not applicable in many real-world scenarios where one has an access to datasets consisting of multiple views/representations (e.g. various omics profiles, medical text records coupled with FMRI images, etc). Proposed method can leverage information from these different views and produce more robust and accurate results in comparison to traditional methods.

KeCo: kernel-based online co-agreement algorithm. This online algorithm uses a co-agreement strategy to take into account unlabelled data and to improve classification performance. Unlike the standard online methods it is naturally applicable to many real-world situations where data is available in multiple representations. In addition, our online algorithm allows learning non-linear relations in the data via kernel functions.

Personalized microbial network inference via co-regularized spectral clustering. Based on the results of co-regularized spectral clustering this code visualizes two groups of individuals with different topology of their microbial interaction network. The results of microbial network inference suggest that niche-wise interactions are different in these two groups. The network visualization is implemented in Python and in Matlab.

Online co-regularized algorithm. The proposed algorithm is particularly applicable to learning tasks where large amounts of (unlabeled) data are available for training. The algorithm co-regularizes prediction functions on unlabeled data points and leads to improved performance in comparison to several baseline methods on UCI benchmarks and a real world natural language processing datasets.

Probabilistic preference learner/ranker - ProbRank. The algorithm can learn a ranking function based on pairwise comparison data, that is, data about the ranking function values is provided in terms of pairwise comparisons at the given locations. This is accomplished in two ways: a) Approximating the marginal likelihood using expectation propagation and carrying out maximum likelihood procedure on the hyper-parameters. In this case the square exponential covariance function is used. b) Considering ranking as a regression with Gaussian noise and Gaussian processes prior, given the score differences.

Joint work with Botond Cseke. Download Matlab implementation of the probabilistic preference learning models described in "Kernel principal component ranking: Robust ranking on noisy data".

E-MaLeS 1.0 (3rd place in the FOF division of the CADE ATP System Competition). E-MaLeS 1.0 is an automated theorem prover which is based on E prover. E-MaleS 1.0 uses E with different strategies than the standard auto mode. Furthermore it employs strategy splitting, e.g. it runs several strategies. Note that this version is very CASC focused.

Joint work with Daniel Kuehlwein. Download the code.

Multi-output ranker for automated reasoning. Joint work with Daniel Kuehlwein. Download the code and the data used in the experiments of the paper "Multi-Output Ranking for Automated Reasoning".


Recent Publications



2018

Sultan Imangaliyev, Andei Prodan, Max Nieuwdorp, Albert K. Groen, Natal van Riel, and Evgeni Levin
Domain Intelligible Models
Methods, 2018 [ html ]

Mélanie Deschasaux, Kristien E. Bouter, Andrei Prodan, Evgeni Levin, Albert K. Groen, Hilde Herrema, Valentina Tremaroli, Guido J. Bakker, Ilias Attaye, Sara-Joan Pinto-Sietsma, Daniel H. van Raalte, Marieke B. Snijder, Mary Nicolaou, Ron Peters, Aeilko H. Zwinderman, Fredrik Bäckhed and Max Nieuwdorp
Depicting the composition of gut microbiota in a population with varied ethnic origins but shared geography
Nature Medicine, 2018 [ html ]

Joao Pereira, Michiel J. Bom, Albert K. Groen, Erik S. G. Stroes, Paul Knaapen, and Evgeni Levin
Protein Space Embedding Kernel for Plaque Volume Prediction
Joint ICML and IJCAI - Workshop on Computational Biology, 2018 [ html ]

Annieke van Baar, Andrei Prodan, Camilla D. Wahlgren, Steen S. Poulsen, Filip K Knop, Albert K. Groen, Jacques J. Bergman, Max Nieuwdorp and Evgeni Levin
Duodenal L cell density correlates with features of metabolic syndrome and plasma metabolites
Endocrine Connections, 2018 [ html ]

Loek P. Smits, Ruud S. Kootte, Evgeni Levin, et al, Stanley L. Hazen and Max Nieuwdorp
Effect of Vegan Fecal Microbiota Transplantation on Carnitine and Choline-Derived Trimethylamine-N-Oxide Production and Vascular Inflammation in Patients With Metabolic Syndrome
Journal of the American Heart Association, 2018 [ html ]

Karin Fieten, Joan Totte, Evgeni Levin, Marta Reyman, Yolanda Meijer, André Knulst, Frank Schuren, Suzanne Pasmans
Fecal Microbiome and Food Allergy in Pediatric Atopic Dermatitis: A Cross-Sectional Pilot Study
Allergy and Immunology, 2018 [ html ]

2017

Evgeni Levin, Ruud Kootte, Albert K. Groen, and Max Nieuwdorp
Learning Microbial Response.
Machine Learning in Computational Biology (NIPS workshop), 2017.


Ruud S. Kootte, Evgeni Levin, Jarkko Salojärvi, Loek P. Smits, Annick V. Hartstra, Shanti D. Udayappan, Gerben Hermes, Kristien E. Bouter, Annefleur M. Koopen, Jens J. Holst, Filip K. Knop, Ellen E. Blaak, Jing Zhao, Hauke Smidt, Amy C. Harms, Thomas Hankemeijer, Jacques J.G.H.M. Bergman, Hans A. Romijn, Frank G. Schaap, Steven W.M. Olde Damink, Mariette T. Ackermans, Geesje M. Dallinga-Thie, Erwin Zoetendal, Willem M. de Vos, Mireille J. Serlie, Erik S.G. Stroes, Albert K. Groen, Max Nieuwdorp
Improvement of Insulin Sensitivity after Lean Donor Feces in Metabolic Syndrome Is Driven by Baseline Intestinal Microbiota Composition.
Cell Metabolism, 2017 [ html ]

Sultan Imangaliyev and Evgeni Levin
Unsupervised Multi-View Feature Selection for Tumor Subtype Identification.
ACM Conference on Bioinformatics, Computational Biology, 2017 [ bib ]

Sultan Imangaliyev, Johannes Matse, Jan Bolscher, Ruud Brakenhoff, David Wong, Elisabeth Bloemena, Enno Veerman and Evgeni Levin
Discovery of Salivary Gland Tumors' Biomarkers via Co-Regularized Sparse-Group Lasso.
Algorithmic Learning Theory/Discovery Science, 2017. [ bib ]

Sara Botschuijver, Guus Roeselers, Evgeni Levin, Daisy Jonkers, Olaf Welting, Sigrid E. Heinsbroek, Hellen H. de Weerd, Teun Boekhout, Matteo Fornai, Ad Masclee, Frank HJ Schuren, Wouter de Jonge, Jurgen Seppen, Rene M. van den Wijngaard
Intestinal Fungal Dysbiosis Associates With Visceral Hypersensitivity in Patients With Irritable Bowel Syndrome and Rats.
Gastroenterology, 2017 [ bib ]

Egija Zaura, Bernd W Brandt, Andrei Prodan, Maarten Joost Teixeira de Mattos, Sultan Imangaliyev, Jolanda Kool, Mark J Buijs, Ferry LPW Jagers, Nienke L Hennequin-Hoenderdos, Dagmar E Slot, Elena A Nicu, Maxim D Lagerweij, Marleen M Janus, Marcela M Fernandez-Gutierrez, Evgeni Levin, Bastiaan P Krom, Henk S Brand, Enno CI Veerman, Michiel Kleerebezem, Bruno G Loos, G A van der Weijden, Wim Crielaard and Bart JF Keijser
On the ecosystemic network of saliva in healthy young adults.
ISME Journal, 2017 [ bib ]

Erwin M. Berendsen, Evgeni Levin, René Braakman, Debora van der Riet-van Oeveren, Norbert Sedee, Armand Paauw
Identification of microorganisms grown in blood culture flasks using Liquid Chromatography-Tandem Mass Spectrometry.
Future Microbiology, 2017 [ bib ]

Sultan Imangaliyev, Monique H. van der Veen, Catherine M. C. Volgenant, Bruno G. Loos, Bart J. F. Keijser, Wim Crielaard, Evgeni Levin
Classification of Quantitative Light-Induced Fluorescence Images Using Convolutional Neural Network.
International Conference on Artificial Neural Networks, 2017 [ bib ]

2016

Ruud Kootte, Evgeni Levin, et al
Predicting response in insulin sensitivity upon lean donor fecal transplantation in metabolic syndrome subjects by identifying microbial biomarkers.
Annual Dutch Diabetes Research Meeting, 2016. [ bib ]

Andrei Prodan, Sultan Imangaliyev, Henk S. Brand, Martijn N. A. Rosema, Evgeni Levin, Wim Crielaard Bart J. F. Keijser, and Enno C. I. Veerman
Salivary metabolome and functional biochemistry of systemically healthy young adults.
Metabolomics, 2016. [ bib ]

Guus Roeselers, Jildau Bouwman, Evgeni Levin
The human gut microbiome, diet, and health: “Post hoc non ergo propter hoc”.
Trends in Food Science & Technology, 2016. [ bib ]

Paula L. Amaral Santos, Sultan Imangaliyev, Klamer Schutte and Evgeni Levin
Feature Selection via Co-regularized Sparse-Group Lasso.
Machine learning, Optimization and Big Data, 2016. [ bib ]

Andrei Prodan, Sultan Imangaliyev, Enno Veerman, Evgeni Levin, Bart Keijser, Wim Crielaard.
A Biomarker for Oral Health Resilience.
European Patent Office, 2016. [ bib |

Andrei Prodan, Henk S. Brand, Antoon J. M. Ligtenberg, Sultan Imangaliyev, Evgeni Levin, Fridus van der Weijden, Wim Crielaard, Bart J. F. Keijser, Enno C. I. Veerman.
A study of the variation in the salivary peptide profiles of young healthy adults acquired using MALDI-TOF MS.
PLOS ONE, 2016. [ bib |

Astrid A.T.M. Bosch, Evgeni Levin, Marlies A. van Houten, Raiza Hasrat, Gino Kalkman, Giske Biesbroek, Wouter A.A. de Steenhuijsen Piters, Pieter-Kees C.M. de Groot, Paula Pernet, Bart J.F. Keijser, Elisabeth A.M. Sanders, Debby Bogaert.
Development of Upper Respiratory Tract Microbiota in Infancy is Affected by Mode of Delivery.
EBioMedicine, 2016. [ bib | http ]

Sultan Imangaliyev, Monique van der Veen, Catherine Volgenant, Bart Keijser, Wim Crielaard and Evgeni Levin
Deep Learning for Classication of Dental Plaque Images.
Machine learning, Optimization and Big Data, 2016. [ bib |

2015

Laurens van de Wiel, Tom Heskes, and Evgeni Levin
KeCo: Kernel-based Online Co-agreement Algorithm.
Algorithmic Learning Theory/Discovery Science, 2015. [ bib | pdf ]

Sultan Imangaliyev, Bart Keijser, Wim Crielaard, and Evgeni Levin.
Deep learning of human microbiome in health and disease.
International Human Microbime Congress, 2015. [ bib | http ]

Sultan Imangaliyev, Bart Keijser, Wim Crielaard, and Evgeni Levin.
Personalized microbial network inference via co-regularized spectral clustering.
Methods, 2015. [ bib | http ]