Dear Students,
At this point you will find general information of our professorship which are not directly connected to any course.
If you have any questions, please do not hesitate to contact us. Either by posting in the Q&A forum below or by contacting one of our group members directly:
The Cognitive Systems group holds a regularly doctoral colloquium. There, our internal and external doctoral students present and discuss the current status of their work. Occasionally, we invite external lecturers. We welcome interested students and colleagues (exceptions are marked with 'internal' in the schedule).
When? Wednesdays, 16:00 s.t. - 17:30
Where? KogSys-Lab (WE5/05.013)
CogSys Lab: WE5/05.013
Every Wednesday: 16:00-17:30 Uhr
| Date | Speaker(s) | Topic |
|---|---|---|
| 6.4.23 | ||
| 13.4.23 | Franz Motzkus | How to Doktor*innenarbeit |
| 20.4.23 | Bettina Finzel | Write Reviews: A How To |
| 27.4.23 | ||
| 4.5.23 |
||
| 10.5.23 | Bettina Finzel | Pixy - ein Tool für Reasoning auf Graphen (im Speziellen Neo4j) |
| 17.5.23 | ||
| 24.5.23 | Bettina Finzel | Concept- and Relation-based Explanations |
| 15.6.2023 | Adish Singla | Fakultäts Kolloquium: AI-Driven Educational Technology for Introductory Programming |
| 21.6.2023 |
Open Session | |
| 28.6.2023 | Open Session | |
| Monday 3.7.2023 14:00-15:00 |
Christoffer Löffler | tba |
| Tuesday 4.7.2023 14:00-15:30 | Martin Schmeißer | tba |
| 5.7.2023 | cancelled | |
| 12.7.2023 | Open Session | |
| 19.7.2023 | Prof. Dr. CC Carbon | Verschwörungstheorien und -mythen: Was, woher, warum - was machen sie mit uns und wie gehen wir mit ihnen um, und das in Zeiten von DeepL, ChatGPT, DALL-e und Kolleg*innen? |
| 26.7.2023 | Open Session | |
| 2.8.2023 | Open Session | |
| 9.8.2023 | Open Session | |
| 16.8.2023 | Open Session remote on zoom (NO hybrid meeting) | |
| 23.8.2023 | Zeynep Saribatur (TU Wien) | Foundations for Projecting Away the Irrelevant in ASP Program |
| Monday 28.8.2023 11:00 |
Alexander Hatzold, Master Angewandte Informatik | Semantic and Syntactic Analogical Reasoning in Large Language Models (LLMs): Effects of Different Types of Prompting on the Performance of ChatGPT |
| 30.8.2023 | Open Session | |
| 6.9.2023 | Open Session | |
| 13.9.2023 | Simon Schramm (BMW) | ILP for explainable graph clustering |
| 19.9.2023 | Felix Hempel | Explanatory and Interactive Machine Learning with Counterfactual Explanations for Ordinal Data |
| 20.9.2023 | ||
| 27.9.2023 | ||
| 4.10.2023 | ||
| 9.10.2023 | Richard Nieding | Unsupervised Machine Learning via Feature Extraction and Clustering to Classify Tree Species from High-Resolution UAV-based RGB Image Data |
| 11.10.2023 | ||
| 18.10.2023 | ||
| 25.10.2023 | ||
| Date | Topic |
|---|---|
| 05.10.2022 (10:00) | Louisa Heidrich: Combining Explanatory Interactive Learning and Fair Machine Learning – Interacting with Explanations to Uncover and Reduce Human and Machine Biases |
| 25.10.2022 (14:00) |
Amit Singh: Does Contrastive Attention Guidance Facilitate Action Recall? An Eye-
tracking Study |
| 10.11.2022 (12:15) |
Zeynep Saribatur: Towards Comprehensible ASP Reasoning by Means of Abstraction |
| 14.11.2022 (16:00) |
David Tafler: The Object-Centric Concept Learner - An Interpretable Framework for Learning Relational Visual Concepts |
| 20.12.2022 | Thomas Sedlmeir: Betrugserkennung in Dokumenten: Untersuchung von Methoden des Maschinellen Lernens für die Identifikation manipulierter Zeichen |
| 19.1.2023 |
Stefan Raab: Active Learning for interactive labeling of RGB UAV forest data Johannes Rabold & Johannes Langer: A Brief Overview of Transformer Models and their Architecture (PDF zum Vortrag) |
| 26.1.2023 |
Open Session |
| 2.2.2023 (tba) |
Maximilian Rosilius: Fehlerauswirkungen in der Mensch-Maschine-Interaktion auf die menschliche Leistung und Akzeptanz |
| 16.2.2023 (18:00-19:00) |
Franz Motzkus: Quantifizierung von lokalen Heatmapping xAI Methoden |
| 02.3.2023 |
Christoph Wehner: Docker Einführung |
| 09.3.2023 |
Mareike Müller: Masterarbeit Verteidigung zum Thema "Dynamic Fingerspelling Recognition Using Sequential Deep Learning" |
| Date | Topic |
|---|---|
| 13.04.2022 (17:00) | Mareike Müller: Automatic Sign Language Recognition |
| 12.05.2022 |
Matthias Uhl: People Prefer Moral Discretion to Algorithms |
| 23.06.2022 | Alexander Becker: Certified Data Removal in Sum Product Networks |
| 05.07.2022 | Felix Schweinfest: Comparing the Performance of Memory Augmented Neural Networks in Reinforcement Learning |
| Date | Topic/Doctoral Candidate |
|---|---|
| 07.10.2021 | Simon Kuhn: Verifying Convolutional Neural Networks: A Landmark-based Approach for Quantitative Evaluation of Facial Action Unit Classification |
| 29.10.2021 | Christoph Wehner: Master's Defense with the topic: "Interpretable Lane Change Prediction" |
| 25.11.2021 |
Oliver Mey: Prediction of Energy Consumption for Variable Customer Portfolios Including Aleatoric Uncertainty Estimation
|
| 02.12.2021 |
Sarem Seitz: Gaussian Processes come with (almost) free explainability |
| 07.12.2021 (17:00!!) |
Pascal Hitzler: Concept Induction for Explainable AI
|
| 13.01.2022 |
|
| 16.02.2022 (17:00!!) |
Talk Aliya Hammad: Optimizing Hierarchical Classification for Resource-Constrained Systems Using Reinforcement Learning Methods. |
| 24.03.2022 |
Talk Christoph Wehner: Root-Cause Detection in the Wild - The first steps of the project "KIProQua" |
| Date | Topic/Doctoral Candidate |
|---|---|
| 15.04.2021 | Bettina Finzel: Interactive Machine Learning |
| 29.04.2021 | Jan Tinapp: Debugging Causal Models. Later: Johannes Rabold: Discussion Research Proposition (Humanly plausible feature vector clustering) |
| 06.05.2021 | Lisa Lengenfelder: Are Relations Relevant in CNNs? |
| 20.05.2021 |
Talk 2: Deepika Arneja: TraMeExCo Internship Presentation (Join via Teams here) |
| 20.05.2021 | Talk 1: Joscha Eirich, Jakob Bonart: ”The interactive clustering and labeling of acoustic signature patterns of electrical engines (We use the same Teams link as one row above!) |
| 27.05.2021 | Oliver Mey: Self-Masking Adversarial Networks |
| 10.06.2021 | Dennis Müller: Mathematical Knowledge Management Across Formal Libraries |
| 24.06.2021 | Wolfgang Kratsch (Credium): Fusion von unterschiedlichen Datenquellen für semi-automatisiertes Labelling und Closed-Loop Machine Learning Cycle |
| 01.07.2021 |
Alexander Schneider: How to Train a NER Model from Annotation to Model |
| 05.07.2021 |
Ploy Schneider: Counterfactuals as an Approach to Test Fairness in Machine Learning -- And Why Fairness in Machine Learning Affects Everyone |
| 15.07.2021 |
Christian Wirth: Scalable Probabilistic Deep Learning: An Engineers Guide to the State of the Art |
| 23.09.2021 |
Maximilian Muschalik: Beyond Contrastive Explanations for Deep Learning Models in Physical Relational Domains
|
| Date | Topic/Doctoral Candidate |
|---|---|
| 01.10.2020 & 08.10.2020 |
KI2020 Recap |
| 15.10.2020 |
AI Safety, Gesina Schwalbe |
| 29.10.2020 | Mena Leemhuis: Cones, Negations and all that |
| 5.11.2020 (will start at 17:15 this time only!) | Durgesh Nandini: Answer Set Programming |
| 12.11.2020 |
Deep Transfer Learning for Time Series Regression in Chemical Production, Namrata Jain (Master Thesis at Wacker AG, Intermediate Presentation) |
| 19.11.2020 |
Deniz Neufeld: Recurrent Neural Network Architecture based on Dynamic Systems Theory for Data Driven Modelling of Complex Physical Systems |
| 26.11.2020 | Sarem Seitz: Bayesian mixtures of interpretable experts |
| 10.12.2020 | Bettina Finzel: Sustainability in Machine Learning: From Nomads to Data Centers and Back. |
| 14.12.2020 | Regina Siegers (Master Survey Statistics, Uni Bamberg): Vectors and the Search for Happiness. |
| 21.01.2021 | Tobias Huber (Uni Augsburg): "Benchmarking Perturbation-based Saliency Maps for Explaining Deep Reinforcement Learning Agents" and "Do Deep Neural Networks Forget Facial Action Units? - Exploring the Effects ofTransfer Learning in Health Related Facial Expression Recognition" |
| 11.03.2021 | mgm: AI assisted labelling tool |
| 18.03.2021 | Lennart Thamm: Intermediate Presentation of thesis results (delayed and will start at 17:00) |
| 25.03.2021 | Gesina Schwalbe: Concept Analysis for Verification -
slides |
| 15.04.2021 | Bettina Finzel: Interactive Machine Learning: Approaches and Examples (this time only click here to join the Teams meeting) |
| Date | Topic/Doctoral Candidate |
|---|---|
| 09.07.2020 |
Joscha Eirich |
| 23.07.2020 | Gesina Schwalbe and Johannes Rabold |
| 17.09.2020 |
Kick-Off in Preparation of Winter Term |
| Date | Topic/Doctoral Candidate |
|---|---|
| 2020-01-30 | Sarem Seitz: "End-to-end prototype-based reinforcement learning" |
| 2020-02-13 | Tobias Hepp: "Latent Class Distributional Regression" |
| 2020-03-19 |
Regina Siegers: "Parallelized Causal Inference with Latent Variables" |
| Date | Topic / Doctoral Candidate |
|---|---|
| 2019-04-18 | Gast: Thomas Voit, TH Nürnberg, Lernen über Spiele/Ähnlichkeit von Spielen |
| 2019-04-25 | |
| 2019-05-02 | internal meeting |
| 2019-05-09 | no meeting |
| 2019-05-16 | no meeting |
| 2019-05-23 | Thoughts on automatic constraint generation from examples (Michael Siebers) |
| 2019-05-30 | public holiday |
| 2019-06-06 | 14-15 Malte Büttner -- Iterative
visual reasoning beyond convolutions |
| 2019-06-13 | Reserved - topic will be announced |
| 2019-06-20 | public holiday |
| 2019-06-27 | no meeting |
| 2019-07-04 | Towards the Inference of Shape
Predicates from Memory Graphs (Jan Boockmann) |
| 2019-07-11 | Making Reinforcement Learning agents efficient for learning and acting in the real-world (Sarem Seitz)(17:15) |
| 2019-07-18 | no meeting |
| 2019-07-25 | Pepper (Sebastian Seufert, Ferdinand Lang) |
| 2019-08-01 | no meeting |
| 2019-08-08 | public holiday |
| 2019-08-15 | public holiday |
| 2019-08-22 | no meeting |
| 2019-08-29 | Der Einfluss von Erklärungen auf Vertrauen in Dare2Del - Studienvorstellung (Clemens Bartnik) |
| 2019-09-05 | |
| 2019-09-12 | external talk (Tim Grelka) 13-14 Uhr! |
| 2019-09-19 | no meeting |
| 2019-09-24, 9:15-11:00 (Tuesday) |
Optimizing alert timings (external talk, Thilip) |
| 2019-09-26 | no meeting |
| Date | Topic / Doctoral Candidate |
|---|---|
| 2018-11-22, 14:00-15:00 | Mark Gromowski internal |
| 2018-11-29 |
Explanation Generation Frameworks |
| 2018-12-06 | Bayesian Deep Learning (Johannes Rabold) |
| 2018-12-13 | ibindo – ein Chatbot gegen Liebeskummer (external talk) [abstract] |
| 2018-12-20 | Gesina Schwalbe, Absicherung von DNNs im autonomen Fahren; afterwards Christmas party |
| 2019-01-10 | cancelled! |
| 2019-01-17 | |
| 2019-01-24 | no meeting |
| 2019-01-31 | Hannah Deininger, Masterarbeit zur expliziten Erklärungsfindung für komplexe Neuronale Netze |
| 2019-02-07 | cancelled! |
| 2019-02-14, 17:00-18:00 |
Bettina Finzel, Constraint-based Explanations in Aleph |
| 2019-02-20 |
Mark Gromowski, Projektupdate PainFaceReader (Schwerpunkt Datenrepräsentation) |
| 2019-02-28 | no meeting |
| 2019-03-07 | ECDA 2019 - Testvorträge / Fragen |
| 2019-03-14 |
Gast: Martin Hillebrand (Vodafone) |
| 2019-03-21 |
Abschlussbericht Praktikant Sebastian Gary |
| 2019-03-28 |
no meeting |
In the Cognitive Systems group, topics for projects and bachelor's and master's theses are usually closely linked to our research. If you like to do your thesis or a project with us, please select a suitable topic from the list below and contact the named contact person. Alternatively, we encourage students to suggest their own topic. In that case, please contact Prof. Schmid or Johannes Rabold. In either case, we expect that you have shown interest in our group, for example by participating in our courses.
Please note that we only offer projects in the summer semesters! These are regular projects for groups of students under one selected topic. For more information see the respective VC courses.
Type: master's thesis
Suggested Knowledge / Courses: depends on topic
Contact: Ute Schmid
Type: master's thesis
Suggested Knowledge / Courses: Machine Learning
Contact: Johannes Rabold or Ute Schmid
For more information see https://vc.uni-bamberg.de/pluginfile.php/1106543/mod_label/intro/2018-11-14_Ausschreibung_Masterarbeit_MEXRT.pdf (German only).
Type: master's thesis / project (15 ECTS)
Suggested Knowledge / Courses: Basic Inductive Logic Programming (e.g. FOIL), Prolog
Contact: Michael Siebers
Inductive Logic Programming (ILP) is a special case of machine learning which aims to learn logic formulas (rules) from examples. For instance, one such rule may be that file A can be deleted if there is another file B in the same directory which has the same size and is older. One shortcoming of ILP methods is the handling of noise in data since (first-order logic) rules can only be true or false, nothing in between. Thus, when applied to new data, rules either classify the instance as 100%-true or 100%-false. However, many classification algorithms (e.g. SVMs, Decision Trees, Naive Bayes) additionally provide a confidence, that is how likely is it that the prediction is true. Aim of this thesis/project is to provide such information for rules learned with Aleph which is a widely used ILP system which uses Prolog to represent examples, background knowledge, and resulting rules.
How Close Do LIME Generated Visual Explanations Match Those of Humans
BA or MA in Psychology
The XAI approach LIME generates visual explanations based on identifying super-pixels in images which contribute significantly to a classification decision. We want to investigate how close those explanations match with what areas in an image humans are highlighting. The thesis builts upon the thesis of Ludwig Schallner who could show that the selected approach to generate super-pixels has an impact on how close LIME explanations correspond to a human expert on simple images. In the planned experiment we want to focus on more complex natural categories (e.g. dog) and also study the impact of the context (dog vs cat; dog vs. bear; dog vs. car; etc.).
(genauere Information bei Ute Schmid erfragen)
Sequence Mining
Die Datenbasis sind typischerweise 1.000 Systeme die täglich jeweils 50.000 Zeilen Event-Log schicken und ca. alle 3 Jahre fällt ein System aus. Unser Ziel ist es in den Event-Logs Sequenzen von Ereignissen zu finden, die darauf hindeuten, dass ein System bald kaputt ist.
Logdata Parsing
Aktuell bekommen wir ca. 100 verschiedene Logfiles, die meisten als csv, manche xml. Die sind teilweise ganz chaotisch, oft muss man als ersten Schritt den richtigen Zeilenumbruch finden und im nächsten Schritt sollen dann die Felder in jeder Zeile getrennt werden. Ca. alle 3 Monate gibt es eine neue Softwareversion, dabei ‚kann‘ sich das Format ändern. Die Frage ist, ob man das Parsing automatisieren kann und auch automatisch erkennen kann, wann sich das Format geändert hat – dann sollten immer automatisch die richtigen Parsing-Regeln verwendet werden.
Optimizing alert timings - Predicting system failures daily vs. raising alerts
Bei unseren Predictive Maintenance
Modellen aggregieren wir typischerweise Daten auf Tagesebene und machen
auch täglich scorings und machen dabei für jeden Tag eine Vorhersage, ob
der Tag in den 30 Tagen vor einem Systemausfall
ist. Aber für unseren use-case brauchen wir gar nicht täglich eine
Wahrscheinlichkeit, da würde es uns reichen wenn innerhalb von dem 30
Tage Fenster mind. 1 Alarm generiert wird – der dann aber möglichst
genau sein sollte, genauer als bei den Betrachtungen
von nur 1 Tag.
Das aktuelle Template für Abschlussarbeiten am Lehrstuhl finden Sie im Git unter folgendem Link (bitte beachten Sie die README): https://gitlab.rz.uni-bamberg.de/cogsys/public/students/thesis-template
Unsere alten Templates finden Sie hier: