Verbundprojekte

Mit dem Format der Verbundprojekte wollen wir gezielt Forschungsschwerpunkte setzen, in denen die Entwicklung und Anwendung von Methoden des maschinellen Lernens zu spezifischen Fragestellungen aus unterschiedlichen wissenschaftlichen Disziplinen vorangetrieben werden. Ziel dieses Formats ist der vertiefte Austausch zwischen den verschiedenen Disziplinen, die sich in einer Art Mini-Graduiertenschule zusammenschliessen. Unsere derzeit 5 geförderten Verbundprojekte bestehen aus jeweils 4 Subrojekten (SP), in denen verschiedene Arbeitsgruppen gemeinsam an folgenden, übergeordneten Themen arbeiten:

Compositionality in Minds and Machines
Machine Learning in Education
Modeling and Understanding Spatiotemporal Environmental Interactions (MUSTEIN)
Probabilistic Inference in Mechanistic Models (PIMMS)
Uncovering the inner structure of medical images through generative modeling

► Compositionality in Minds and Machines

Compositionality makes all interesting computation possible. By combining and recombining smaller parts into a meaningful whole, compositionality boosts generalization, reduces sample complexity, and improves interpretability. From convolutional networks to structured probabilistic inference, many frameworks can be modelled as compositional functions for pattern recognition, reasoning, and planning. Indeed, inferring good models from limited training sets — and all training sets are limited — fundamentally requires compositionality.

In this Network Project, across 4 Subprojects (SP) we bridge different disciplines to create models of compositional reasoning in minds and machines: In SP 1, we work towards a model that can perform common-sense compositional zero-shot generalization in open world scenarios. In SP 2, we build models that can reason compositionally to explore their environment and solve abstract intelligence tasks in a human-like fashion. In SP 3, we use the compositional nature of human gaze to improve fine-grained classification. In SP 4, we study which inductive biases allow humans to learn and compositionally reuse latent structures.

SP 1
Compositional Zero-Shot Learning, Disentangling Representations and Generalization

  •  Projektleiter:              Zeynep Akata1 · zeynep.akata@uni-tuebingen.de · eml-unitue.de
                                          Eric Schulz2 · eric.schulz@tuebingen.mpg.de · cpilab.org
                                          Matthias Bethge3,4 · matthias@bethgelab.org, bethgelab.org
     
  •  Projektmitarbeiter:   Shyamgopal Karthik1 (Doktorand)

                                                       Universität Tübingen, 1Exzellenzcluster Maschinelles Lernen, 3Fachbereich Physik, 4Tübingen AI Center,
                                                                      2Max Planck Institut für Biologische Kybernetik, Tübingen

 Projektlaufzeit: Oktober 2021 - Februar 2025

 

SP 2
Compositional reasoning in combinatorial spaces

  •  Projektleiter:             Eric Schulz1 · eric.schulz@tuebingen.mpg.de · cpilab.org
                                         Zeynep Akata2 · zeynep.akata@uni-tuebingen.de · eml-unitue.de
                                         Peter Dayan1 · dayan@tue.mpg.de · kyb.tuebingen.mpg.de/computational-neuroscience
     
  •  Projektmitarbeiter:   Tankred Saanum1 (Doktorand)

                                            1Max Planck Institut für Biologische Kybernetik, Tübingen, 2Universität Tübingen, Exzellenzcluster Maschinelles Lernen
 

 Projektlaufzeit: März 2021 - Februar 2025

 

SP 3
Compositional Representations of Human Gaze

  •  Projektleiter:             Enkelejda Kasneci1 · enkelejda.kasneci@uni-tuebingen.de · www.hci.uni-tuebingen.de
                                         Zeynep Akata2 · zeynep.akata@uni-tuebingen.de · eml-unitue.de
                                         Felix Wichmann1 · felix.wichmann@uni-tuebingen.de · wichmannlab.org 
     
  •  Projektmitarbeiter:   Michael Kirchhoff1 (Doktorand)

                                                Universität Tübingen, 1Fachbereich Informatik, 2Exzellenzcluster Maschinelles Lernen
 

 Projektlaufzeit: Juli 2021 - Dezember 2024

 

SP 4
Inductive Biases in Compositional Sequence Learning

  •  Projektleiter:              Charley Wu1 · charleymswu@gmail.com · hmc-lab.com
                                          Martin Butz2 · martin.butz@uni-tuebingen.de · cm.inf.uni-tuebingen.de
                                          Eric Schulz3 · eric.schulz@tuebingen.mpg.de · cpilab.org
     
  •  Projektmitarbeiter:   Lee Sharkey1 (Doktorand)
                                         
    Turan Orjulu (Doktorand)

                             Universität Tübingen, 1Exzellenzcluster Maschinelles Lernen, 2Fachbereich Informatik,
                                                   3Max Planck Institut für Biologische Kybernetik, Tübingen

 

 Projektlaufzeit: November 2021 - November 2025

 


► Machine Learning in Education

Education is designed to foster learning in complex, real-life environments. Learners, however, differ substantially in many respects, including social, cognitive, and motivational factors, and their prior knowledge in the target domain.

Education needs to take this substantial heterogeneity into account. With the increasing use of Intelligent Tutoring Systems (ITS) in real-life education, using machine learning to optimally support individual learning is becoming feasible. Yet, we still lack the ability to
(1) interpret rich learning process data digitally collected in real-life education contexts in a way that supports an understanding of learning outcomes based on the characteristics of individual learners and learning activities. On this basis, optimally supporting individual learners then requires,
(2) taking the structure of the knowledge domain into account, and
(3) adaptively sequencing activities to best match individual skills and challenges. While ML methods are promising for addressing these gaps, we need to consider
(4) how determining learning opportunities based on ML models can be designed to be fair, and how to handle rich learner and learning-process data in a privacy-respecting way.

In this Network Project, we address these four intertwined research gaps, where interdisciplinary collaboration and the advancement of machine learning methods is crucial, within 4 Subprojects (SP1-SP4 below).

SP 2
Modeling learning in structured domains

  •  Projektleiter:             Álvaro Tejero-Cantero1 · alvaro.tejero@uni-tuebingen.de · mlcolab.org
                                         Charley Wu1 · charley.wu@uni-tuebingen.de · hmc-lab.com
     
  •  Co-Advisors:              Detmar Meurers2, Kou Murayama3, and Ulf Brefeld4
     
  •  Projektmitarbeiter:  Hanqi Zhou1 (Doktorand)

                                                        1Universität Tübingen, Exzellenzcluster Maschinelles Lernen, 2Fachbereich Computerlinguistik,
                                                        3Hector-Institut für Empirische Bildungsforschung, 4Leuphana University, Institute for Information Systems

 Projektlaufzeit: April 2022 - März 2025

SP 3
Adaptivity and personalized learning in the flow

  •  Projektleiter:             Enkelejda Kasneci1 · enkelejda.kasneci@uni-tuebingen.de · www.hci.uni-tuebingen.de
                                         Detmar Meurers2 · dm@sfs.uni-tuebingen.de · sfs.uni-tuebingen.de/
     
  •  Projektmitarbeiter:  Efe Bozkir1 (Doktorand)

                                                        Universität Tübingen, 1Fachbereich Informatik, 2Seminar für Sprachwissenschaft
 

 Projektlaufzeit: Juli 2021 - Juni 2024

SP 4
Ethics, Privacy, and Fairness in digital education environments

  •  Projektleiter:             Konstantin Genin1 · konstantin.genin@gmail.com · kgenin.github.io
                                         Thomas Grote1 · thomas.grote@uni-tuebingen.de · Webseite
                                         Benjamin Nagengast2 · benjamin.nagengast@uni-tuebingen.de · Webseite
                                         Bob Williamson3 · bob.williamson@uni-tuebingen.de · Webseite
     
  •  Projektmitarbeiter:   Vlasta Sikimic1 (Doktorandin)

                                                        Universität Tübingen, 1Exzellenzcluster Maschinelles Lernen, 2Hector-Institut für Empirische Bildungsforschung,
                                                        3Fachbereich Informatik

 

 Projektlaufzeit: März 2022 - Juni 2023


► Modeling and Understanding Spatiotemporal Environmental Interactions (MUSTEIN)

Our world changes rapidly due to mankind creating the “Anthropocene”. Increasing greenhouse gas concentrations (including CO2 and methane) in the atmosphere lead to climate change. Intensive irrigation, agriculture generally, grubbing, dam building, and impervious surfaces, severely affect our hydrosphere – including our freshwater resources – as well as our biogeosphere – including unprecedented rates of soil erosion. Meanwhile, humans demand more energy in their anthroposphere, while relying on safe and rewarding habitats as well as on sufficient food and water resources.

The Network Project MUSTEIN aims at developing machine learning techniques that reliably learn explainable models of critical aspects of these four highly interacting spheres, focusing in 4 Subprojects (SP) on seasonal weather dynamics (SP1), river water discharge (SP2), soil erosion (SP3), and solar thermal systems (SP4).

SP 1
Seasonal weather forecast

  •  Projektleiter:             Bedartha Goswami1 · bedartha.goswami@uni-tuebingen.de · website
                                         Martin Butz2 · martin.butz@uni-tuebingen.de · https://cm.inf.uni-tuebingen.de/
                                         Hendrik Lensch2 · hendrik.lensch@uni-tuebingen.de · https://uni-tuebingen.de/de/26024
                                         Nicole Ludwig4 · nicole.ludwig@uni-tuebingen.de · https://nicoleludwig.github.io/
     
  •  Projektmitarbeiter:    Jannik Thuemmel1 (Doktorand)

                                                        Universität Tübingen, 1Exzellenzcluster Maschinelles Lernen, 2Fachbereich Informatik, 3ML in Nachhaltigen
                                                                           Energiesystemen

 

 Projektlaufzeit: September 2021 - August 2025

SP 2
Water discharge prediction

  •  Projektleiter:             Martin Butz1 · martin.butz@uni-tuebingen.de · https://uni-tuebingen.de/de/25369
                                         Christiane Zarfl2 · christiane@zarfl@uni-tuebingen.de · https://uni-tuebingen.de/de/84458
  •  Projektmitarbeiter:  Fedor Scholz1 (Doktorand)

                                                        Universität Tübingen, 1Fachbereich Informatik, 2Fachbereich Geowissenschaften
 

 Projektlaufzeit: September 2021 - August 2025

SP 3
Modeling soil erosion

  •  Projektleiter:             Thomas Scholten1 · thomas.scholten@uni-tuebingen.de · website
                                         Martin Butz2 · martin.butz@uni-tuebingen.de · https://uni-tuebingen.de/de/25369
                                         Georg Martius3 · georg.martius@tuebingen.mpg.de · http://al.is.mpg.de
     
  •  Projektmitarbeiter:  Manuel Traub1 (Doktorand)

                                                        Universität Tübingen, 1Fachbereich Geowissenschaften, 2Fachbereich Informatik, 3Max Planck
                                                                           Institut für Intelligente Systeme, Tübingen

 

 Projektlaufzeit: April 2021 - März 2025

SP 4
Solar thermal systems

  •  Projektleiter:             Volker Franz1 · volker.franz@uni-tuebingen.de · website
                                         Carl-Johann Simon-Gabriel2 · cjsg@ethz.ch · website
                                         Georg Martius3 · georg.martius@tuebingen.mpg.de · http://al.is.mpg.de/
                                         Nicole Ludwig4 · nicole.ludwig@uni-tuebingen.de · https://nicoleludwig.github.io
     
  •  Projektmitarbeiter:  Florian Ebmeier1 (Doktorand)

                                                        Universität Tübingen, 1Fachbereich Informatik, 4Exzellenzcluster Maschinelles Lernen,
                                                                          2Learning & Adaptive Systems, ETH Zürich, 3Max Planck Institut für Intelligente Systeme, Tübingen

------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Aim of the project

Providing heating and hot water for buildings contributes considerably to the carbon dioxide emissions in Germany. Solar thermal systems could substantially reduce these emissions. However, the use of such systems is wanting and has even been in decline, which seems partially due to problems in quality control and efficiency. We aim to improve solar thermal systems using machine learning techniques that employ spatiotemporal sensor data as well as meteorological data.


Probabilistic Inference in Mechanistic Models (PIMMS)

Numerical models are at the heart of scientific understanding in many disciplines. In geoscience, numerical models are used to describe geophysical processes; in neuroscience, they provide a testbed of our mechanistic understanding of physiological processes; and in systems cell biology they are used to describe regulatory networks of genes or proteins. Typically such mechanistic models come as systems of differential equations, be it ordinary differential equations (ODEs), simulating a spatial process through time, or partial differential equations (PDEs), simulating a process through both time and space.

In 4 Subprojects (see below), we want to apply currently available probabilistic numerical tools to difficult problems in these three scientific disciplines where ODE/PDE models are paramount. Through these diverse scientific application fields, our primary objective is the development of a general toolbox for probabilistic inference techniques for mechanistic models that will be applicable across disciplines for decades to come.

SP 1
Inferring landscape development processes

  •  Projektleiter:             Todd Ehlers1 · todd.ehlers@uni-tuebingen.de · esdynamics.net
                                         Philipp Hennig2 · philipp.henig@uni-tuebingen.de · mml.inf.uni-tuebingen.de
     
  •  Projektmitarbeiter:   Schmidt Jonathan2 (Doktorand)

                                                        Universität Tübingen, 1Fachbereich Geowissenschaften, 2Fachbereich Informatik
 

 Projektlaufzeit: Januar 2022 - Dezember 2025

SP 2
Unravelling the ice dynamic, atmospheric and oceanographic history of ice-sheets with Bayesian inference

  •  Projektleiter:             Reinhard Drews1 · reinhard.drews@uni-tuebingen.de · website
                                         Jakob Macke2 · jakob.macke@uni-tuebingen.de · uni-tuebingen.de/en/196970
     
  •  Projektmitarbeiter:  Guy Moss2 (Doktorand)

                                                        Universität Tübingen, 1Fachbereich Geowissenschaften, 2Fachbereich Informatik
 

 Projektlaufzeit: Oktober 2021 - September 2025

SP 3
Efficient inference and simulation for mechanistic models in neuroscience

  •  Projektleiter:             Philipp Berens1 · philipp.berens@uni-tuebingen.de · berenslab.org
                                         Jakob Macke2 · jakob.macke@uni-tuebingen.de · uni-tuebingen.de/en/196970
                                         Philipp Hennig3 · philipp.hennig@uni-tuebingen.de · mml.inf.uni-tuebingen.de
     
  •  Projektmitarbeiter:  Jonas Beck1 (Doktorand)

                                                        1Forschungsinstitut für Augenheilkunde, Universitätsklinikum Tübingen
                                                        Universität Tübingen, 2Exzellenzcluster Maschinelles Lernen, 3Fachbereich Informatik

 

 Projektlaufzeit: Januar 2022 - Dezember 2024

SP 4
Inference of predictive genome wide mechanistic models of cellular differentiation from single-cell transcriptomics with gradient information

  •  Projektleiter:             Manfred Claassen1 · manfred.claassen@uni-tuebingen.de ·
                                         Jakob Macke2 · jakob.macke@uni-tuebingen.de · uni-tuebingen.de/en/196970
     
  •  Projektmitarbeiter:  Sebastian Bischoff1 (Doktorand)

                                                        1Innere Medizin I, Universitätsklinikum Tübingen, 2Exzellenzcluster Maschinelles Lernen, Universität Tübingen
 

 Projektlaufzeit: Oktober 2021 - September 2025


Uncovering the inner structure of medical images through generative modeling

Automated analysis of medical image data has undergone impressive developments over the past decade mainly due to the application of deep learning methods for image classification, segmentation and image-based regression. Despite these advances, machine learning (ML) methods are still rarely used in clinical practice due to underperformance in real-life settings.

In this Network Project we want to explore how state-of-the art ML methodology can be applied to medical image data in order to (1) better understand the structure and content of medical images and thus the underlying physiological and pathological processes, (2) assess robustness and safety of ML models for medical image analysis, and (3) increase robustness of ML models for medical image analysis.

Within 4 Subprojects (SP) we aim at developing probabilistic generative models as well as incorporating determiniscially known factors such as the image generation process where such information is available. In this manner, we will obtain useful data representations for medical image analysis and simplify the modelling process through prior knowledge of deterministic mechanisms. From a medical point of view, we will target different medical image data covering 2D, 3D and 3D+t scenarios including normal and pathological cases.

SP1
Domain Generalisation for Medical Image Analysis via Causal Representation Learning

  •  Projektleiter:             Bernhard Schölkopf1· bernhard.schoelkopf@tuebingen.mpg.de · Webseite
                                         Sergios Gatidis1· sergios.gatidis@med.uni-tuebingen.de · Webseite
     
  •  Projektmitarbeiter:   Sergios Gatidis1
                                                
    Erick Michael Cobos Tandezo

                                                      1Max Planck Institute für Intelligente Systeme, Tübingen, 2Radiologie, Universitätsklinikum Tübingen
     

 Projektlaufzeit: Januar 2022 - September 2023

SP 2
Assessing the robustness of medical prediction algorithms

  •  Projektleiter:             Christian Baumgartner1· christian.baumgartner@uni-tuebingen.de · Webseite
                                         Matthias Hein2· matthias.hein@uni-tuebingen.de · Webseite
     
  •  Projektmitarbeiter:   Nikolas Morshuis1 (Doktorand)

                                              Universität Tübingen, 1Exzellenzcluster Maschinnelles Lernen, 2Fachbereich Informatik
 

 Projektlaufzeit: Oktober 2021 - September 2025

SP 3
Leveraging anatomical and temporal priors for prediction andunderstanding of disease in high-dimensional medical imaging data

  •  Projektleiter:             Jakob Macke1· jakob.macke@uni-tuebingen.de · Webseite
                                         Christian Baumgartner1· christian.baumgartner@uni-tuebingen.de · Webseite
     
  •  Projektmitarbeiter:    Jaivardhan Kapoor1 (Doktorand)

                                               1Universität Tübingen, Exzellenzcluster Maschinelles Lernen
 

 Projektlaufzeit: September 2021 - August 2025

SP4
Understanding the structure of fundus images in ophthalmologyusing hybrid generative models including the physical imaging process

  •  Projektleiter:             Philipp Berens1· philipp.berens@uni-tuebingen.de · berenslab.org
                                         Hendrik Lensch2· hendrik.lensch@uni-tuebingen.de · Webseite
     
  •  Projektmitarbeiter:  Sarah Müller1 (Doktorandin)

                                              1Universitätsklinikum Tübingen, Forschungsinstitut für Augenheilkunde, 2Universität Tübingen, Fachbereich Informatik
 

 Projektlaufzeit: April 2021 - März 2025