Reseach seminar

Most sessions take place in forms of a reading group or a research presentation by some member of the group or external guest.

When and where

Each Wednesday 11:00 - 12:30, hybrid setting
Offline: Research Institute CODE, Bundeswehr University Munich.
Online: via WebEx. If you are interested, contact Emmanouil Panagiotou for the link.

Who

AIML group’s members and everybody who is interested in AI/ML.

Meetings

Autumn trimester

• 13.12.2023 Prof. Gerhard Wunder from FU Berlin gave a talk on "On Gradient-like Explanation under a Black-box Setting: When Black-box Explanations Become as Good as White-box" at the CODE colloqium Eirini is co-organizing. We had the opportunity to discuss our research and explore potential collaborations.
• 22.11.2023 Prof. Arthur Zimek from SDU Denmark gave a talk on "Fairness in Imbalanced Classification: An Adjustment to the k Nearest Neighbor Classifier" at the CODE colloqium Eirini is co-organizing. We had the opportunity to discuss our research and explore potential collaborations.
• 08.11.2023 Vivek will give a summary of his current research on augmenting knowledge.
• 31.10.2023 Arjun will give a summary of his current research on the landscape of discrimination in the model parameter space.
• 27.09.2023 Valerie Krug, Ph.D. student at the Otto-von-Guericke-University Magdeburg will give a talk on Neuroscience-Inspired Analysis and Visualization of Deep Neural Networks, and in particular, on how to analyze DNNs with an Event-Related Potentials-inspired technique and how to visualize their activations as topographic maps.
• 13.09.2023 Manolis will give a summary of his current research, which includes two parts: i) Learning impartial policies for sequential counterfactual explanations using Deep Reinforcement Learning</i> (to be presented at the DynXAI Workshop, co-located with ECML PKDD 2023) and ii) Generating synthetic design alternatives in mixed feature spaces - Application on offshore substructure data</i> (related to the SFB 1463 project)

Previous sessions

• 04.07.2023 Prof. Kate Smiles will give an invited talk on "Stress-testing Algorithms via Instance Space Analysis".
• 31.05.2023 Invited speaker Tobias will talk about "Compressive Sensing and Geometry of Inverse Problems for Sparse Signals" and Siamak will present his recent accepted paper (EWAF 2023) "Affinity Clustering Framework for Data Debiasing using Pairwise Distribution Discrepancy"
• 24.05.2023 Manolis will present his work on working with mixed data
• 17.05.2023 Swati will present her work "Lifelong News Bias Prediction in low-resource multilingual headlines"
• 10.05.2023 Tai will present his new PAKDD 2023 paper "Multi-fair capacitated students-topics grouping problem"
• 29.03.2023 Vivek Kumar will talk about "Knowledge Augmenting Practices for domain adaptation in language models".
  
• 22.03.2023 Paper discussion: Friedrich, F. et al (2023). Fair Diffusion: Instructing Text-to-Image Generation Models on Fairness.
• 15.03.2023 Vasilis Gkolemis will present his paper "DALE: Differential Accumulated Local Effects for efficient and accurate global explanations", accepted at ACML22.
• 01.03.2023 Paper discussion: Hinton, G. (2022). The forward-forward algorithm: Some preliminary investigations.
• 21.02.2023 Chethan will present his master thesis "Topological Analysis of Curved Spaces and Hybridization of Autoencoders".