Ivana Jovanovic Buha, M.Sc. (hons)
Technical University of Munich
TUM School of CIT
Department of Computer Science
Boltzmannstrasse 3
85748 Garching
Germany
Office: MI 02.05.040
Mail: ivana.jovanovic (at) tum.de
Tel: +49-89-289-18613
Office Hours: by arrangement
Background
- Doctoral candidate, CIT Graduate Center, TUM Graduate School -
- Research Associate (Wissenschaftlicher Mitarbeiter) at TUM SCCS since June 2018
- M.Sc. with Honours in Computational Science and Engineering, Technical University of Munich, 2018
- Diploma in Electrical Engineering, University of Belgrade, Signals and Systems Department, 2014
Research interests
My research focuses on applied and computational mathematics, particularly on uncertainty quantification (UQ), modeling and system identification, inverse problems, and data-driven model learning. The main applications driving my research are from hydrology. More precisely, in my work, I am bridging the gap between theoretical work on High-dimensional Uncertainty Quantification and Bayesian Inversion, applied to relatively simple simulation models, and more complex real-world problems.
- High-dimensional Forward Uncertainty Quantification and Sensitivity Analysis (mainly, analysis of conceptual distributed hydrologic models)
- Sparse Grids Methods
- Inverse problems - Bayesian Inference
- Machine Learning
Publications
Talks
Posters
- Efficient Uncertainty Quantification and Global Time-Varying Sensitivity Analysis Using the Spatially Adaptive Combination Technique. SIAM Conference on Uncertainty Quantification (UQ22), SIAM, 2022Atlanta, Georgia more…
- Efficient Uncertainty Quantification and Global Time-Varying Sensitivity Analysis of Conceptual Hydrological Model. SIAM Conference on Computational Science and Engineering (CSE21), SIAM, 2021Fort Worth, Texas, U.S.A. more…
Open and running student projects
Runnin student projects
- Danylo Movchan: "Large Scale Outdoor Scene Reconstruction with 3D Gaussian Splatting". Master's Thesis, CIT School - Computer Science Department; in collaboration with Stanford University. Since October 2024
- Stefan Stöckl: "Efficient Bayesian Inference of Hydrological Model Parameters: Mathematical Analysis and Implementation of Markov Chain Monte Carlo Approaches". Master's Thesis, CIT School - Computer Science Department; Since December 2024
- Y. R.: "Efficient Uncertainty Analysis Using Sparse Grids and Polynomial Chaos Expansion: Exploring Different Approaches". Bachelor's Thesis, Bachelor Informatik, CIT School - Computer Science Department. Since December 2024
Open student projects
If you are interested in a student project (Bachelor's or Master's Thesis or anything else), it is the best to contact me directly. Here is the list of some projects that I would offer at the moment (the list is not exhaustive):
- "UQ and SA of Hydrologic model HBV using pyApprox software tool"
- Relevant Sources:
- "Running Uncertainty Quantification simulations using the UM-Bridge software tool: application on a hydrologic model"
- Relevant Sources:
You can also come to my office and discuss possible topics.
See also the list of Student Projects at our chair, e.g., https://www.in.tum.de/i05/jobangebote-studentische-projektarbeiten/job-offers-student-projects/uncertainty-quantification/
Do you want to know what other students are working on in our chair? You are warmly encouraged to attend their presentations at the SCCS Colloquium! Come to get ideas, meet your potential supervisor, or to learn from the style of others for your own presentation.
Finished Student Projects
2024
- Accounting for states and parameter uncertainty of the HBV-SASK hydrologic model using particle filtering as a sequential data assimilation technique. Bachelor thesis, 2024 more… BibTeX
- Efficient Bayesian Inference of Hydrological Model Parameters: Implementation of a Parallel Markov Chain Monte Carlo Approach. Bachelor thesis, 2024 more… BibTeX Full text (mediaTUM)
- Applying recurrent neural networks (RNNs) in the field of hydrology to explore uncertainty in time series forecasts and enhance theory-based models. Bachelor thesis, 2024 more… BibTeX Full text (mediaTUM)
- Artificial Intelligence for Team Productivity. Bachelor thesis, 2024 more… BibTeX Full text (mediaTUM)
2022
- Using the Spatially Adaptive Combination Technique for Efficient Quantification of Uncertainty in Hydrological Models. Bachelor thesis, 2022 more… BibTeX Full text (mediaTUM)
2021
- Second-Order Optimization Methods for Bayesian Neural Networks. Master thesis, 2021 more… BibTeX
- Development of the Bayesian Recurrent Neural Network Architectures for Hydrological Time Series Forecasting. Bachelor thesis, 2021 more… BibTeX Full text (mediaTUM)
- Development of Recurrent Neural Network Architectures for Hydrological Time Series Forecasting. Bachelor thesis, 2021 more… BibTeX Full text (mediaTUM)
2020
- Parallel Evaluation of Adaptive Sparse Grids with Application to Uncertainty Quantification of Hydrology Simulations. Projekt thesis, 2020 more… BibTeX Full text (mediaTUM)
- Developing a prototype of Bayesian Inference framework to recalibrate the complex hydrological model LARSIM. Studien thesis, 2020 more… BibTeX Full text (mediaTUM)
2019
- Development of a Prototype to Quantify the Uncertainty of the Water Balance Model LARSIM. Bachelor thesis, 2019 more… BibTeX Full text (mediaTUM)
- Implementation of a deep learning based model for rainfall-runoff modelling. Master thesis, 2019 more… BibTeX
Other Supervised Student Projects
- Leon Fiedler: "Sensitivity Analysis of a Deep Learning Model for Discharge Prediction in the Regen Catchment". Masterarbeit, Ingenieurfakultät Bau Geo Umwelt; 2020 [BibTeX] [Volltext (mediaTUM)],
News
Teaching
Summer semester 2024
- Algorithms for Uncertainty Quantification
Winter semester 2023
Summer semester 2022
- Seminar Data Mining (IN0014, IN4927)
- Seminar High Dimensional Methods in Scientific Computing (IN2107, IN0014, IN218306)
Winter semester 2021/22
- Einführung in die wissenschaftliche Programmierung (IN8008) [TUMonline] (Moodle)
Summer semester 2021
- Seminar Data Mining (IN0014, IN4927)
- Seminar High Dimensional Methods in Scientific Computing (IN2107, IN0014, IN218306)
Winter semester 2020/21
Summer semester 2020
- Seminar Data Mining (IN0014, IN4927)
- Seminar High Dimensional Methods in Scientific Computing (IN2107, IN0014, IN218306)
Winter semester 2019/20
- Scientific Computing 1 (IN2005) (Moodle)
- Seminar Computational Aspects of Machine Learning (IN2107,IN0014, IN2183)
Summer semester 2018/19
Winter semester 2018/19
- Scientific Computing 1 tutorials (Moodle)
- Assisting in the seminar Next Generation High-Performance Computing.
Winter semester 2016/17
- Assisting in the Scientific Computing 1 tutorials