Universität Bamberg
Suche öffnen
Sie befinden sich hier:

Fakultät Wirtschaftsinformatik und Angewandte Informatik

Lehrstuhl für Erklärbares Maschinelles Lernen

  1.  Otto-Friedrich-Universität Bamberg
  2. Fakultäten
  3. Wirtschaftsinformatik und Angewandte Informatik
  4. Professuren und Lehrstühle
  5. Angewandte Informatik
  6. Lehrstuhl für Erklärbares Maschinelles Lernen
  7. Mitarbeiter
  8. Jonas Alle, M.Sc.

Jonas Alle

Jonas Alle

Teaching and Research Assistant

M.Sc., Doctoral Candidate

Anschrift:      An der Weberei 5, 96047 Bamberg
Raum:            WE5/04.090

Email:             jonas.alle(at)uni-bamberg.de

Sprechstunde: ist nach Vereinbarung per Email möglich

Biography:

Jonas Alle is a dedicated PhD candidate at the Otto-Friedrich University of Bamberg, focusing on the field of Explainable Machine Learning (xAI) since 2024. Jonas holds a master’s degree in Computational Engineering with a focus on Deep Learning and Computational Materials Science, which he earned at the Friedrich-Alexander-University Erlangen-Nuremberg. He also completed a bachelor's degree in Applied Mathematics and Physics at the university of applied sciences Georg-Simon-Ohm in Nuremberg, where he focused on Dynamic Systems and Machine Learning.

His current research centers around the investigation of information flow through deep neural networks, using the perspectives of geometry, probability, or mathematical rigor, to ultimately aim for faithful and robust uncertainty estimations with Bayesian Neural Networks.

Prior to his doctoral studies, Jonas gained significant experience as a Student Research Assistant at the Fraunhofer Institute for Integrated Circuits (IIS), Development Centre for X-Ray Technology. There, he developed deep neural networks for semantic- and instance segmentation tasks for reconstructed 3D images of CT-scans. His bachelor’s thesis focused on plant root segmentation using Convolutional Neural Networks, and the development of an adaptive and efficient inference algorithm. During his master’s thesis, he explored uncertainty estimation techniques for deep learning models and integrated them into a dynamic flood-filling framework for instance segmentation of XXL-CT data.

Jonas has engaged in various extracurricular and professional development activities in 2023. He participated in the AI safety bootcamp "Machine Learning for Good" in Paris and followed the eight-week AI Alignment course by BlueDot Impact.  He also attended the Probabilistic Numerics Spring School at Eberhard Karls University in Tübingen.

Outside of academia, Jonas has a passion for music, frequently visiting concerts and discovering new artists across various genres including rock, metal, and folk. He, also enjoys indoor bouldering.

Profiles: LinkedIn, GitHub, ORCID

Since 2024                                       PhD Candidate at the xAI Lab of the Otto-Friedrich University Bamberg, Germany 
2023Probabilistic Numerics Spring School, three day school and research workshop, Eberhard Karls Universität, Tübingen, Germany 
2023Machine Learning for Good (ML4G), ten day AI safety bootcamp, Paris, France 
2019 - 2023Master of Science in Computational Engineering, Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany 
2018 - 2023Student Research Assistant, Development of deep neural networks for segmentation tasks on reconstructed CT-volumes, Fraunhofer IIS, Germany 
2015 - 2019Bachelor of Science in Applied Mathematics and Physics, Technische Hochschule Georg-Simon-Ohm, Nürnberg, Germany 

 

Main Research Interests

  • Faithful and robust uncertainty estimation with Bayesian Neural Networks
  • Probability distributions in input-, output-, embedding-, and parameter space
  • Theoretical investigations of information flow through deep neural networks
  • Mathematical exploration of Random Neural Networks
  • Research in embedding space geometry, topology, and related characteristics
  • Finding theoretical, mathematically founded, guarantees for deep learning models
  • Insightful data visualization

 

  • Alle J.: Uncertainty Estimation for Instance Segmentation of Large-Scale CT Data with Flood-Filling Networks. Master's Thesis, Friedrich-Alexander University Erlangen-Nuremberg. 2023. (unpublished)
  • 3D Segmentation of Plant Root Systems using Spatial Pyramid Pooling and Locally Adaptive Field-of-View Inference: Alle,J., Gruber,R., Wörlein,N., Uhlmann,N., Claußen,J., Wittenberg,T., Gerth,S.; Frontiers in Plant Science, vol. 14, (2023), https://doi.org/10.3389/fpls.2023.1120189

Thesis Supervision

Please check out our official bidding for thesis topics [Link to VC-Course] or contact me directly via email to request supervision of your Bachelor’s or Master’s Thesis.

WS24/25

  • xAI-Proj-B: Bachelor Project Explainable Machine Learning
  • xAI-DL-M: Deep Learning Exercise