Profiles2020-2025
Kristian Kersting
TU Darmstadt
20Papers
16Last author
35Unique co-authors
0First author
0Solo
First and last author come from byline position, which some venues assign alphabetically.
Recent papers20
- xLSTM-Mixer: Multivariate Time Series Forecasting by Mixing via Scalar MemoriesAuthor 4 of 4NeurIPS 2025 · 27 citations
- Exploring Neural Granger Causality with xLSTMs: Unveiling Temporal Dependencies in Complex DataAuthor 3 of 4NeurIPS 2025 · 5 citations
- Finding NeMo: Localizing Neurons Responsible For Memorization in Diffusion ModelsAuthor 3 of 5NeurIPS 2024 · 46 citations
- Be Careful What You Smooth For: Label Smoothing Can Be a Privacy Shield but Also a Catalyst for Model Inversion AttacksAuthor 3 of 3ICLR 2024 · 26 citations
- Learning Large DAGs is Harder than you Think: Many Losses are Minimal for the Wrong DAGAuthor 5 of 5ICLR 2024 · 9 citations
- Graph Neural Networks Need Cluster-Normalize-Activate ModulesAuthor 5 of 5NeurIPS 2024 · 8 citations
- Rickrolling the Artist: Injecting Backdoors into Text Encoders for Text-to-Image SynthesisAuthor 3 of 3ICCV 2023 · 65 citations
- Interpretable and Explainable Logical Policies via Neurally Guided Symbolic AbstractionAuthor 5 of 5NeurIPS 2023 · 64 citations
- ATMAN: Understanding Transformer Predictions Through Memory Efficient Attention ManipulationAuthor 6 of 7NeurIPS 2023 · 45 citations
- Characteristic CircuitsAuthor 3 of 3NeurIPS 2023 · 8 citations
- ILLUME: Rationalizing Vision-Language Models through Human InteractionsAuthor 4 of 4ICML 2023 · 8 citations
- Do Not Marginalize Mechanisms, Rather Consolidate!Author 5 of 5NeurIPS 2023 · 3 citations
- Safe Latent Diffusion: Mitigating Inappropriate Degeneration in Diffusion ModelsAuthor 4 of 6CVPR 2023
- CLEVA-Compass: A Continual Learning Evaluation Assessment Compass to Promote Research Transparency and ComparabilityAuthor 4 of 4ICLR 2022 · 40 citations
- Interactive Disentanglement: Learning Concepts by Interacting with their Prototype RepresentationsAuthor 4 of 4CVPR 2022 · 14 citations
- Right for Better Reasons: Training Differentiable Models by Constraining their Influence FunctionsAuthor 5 of 5AAAI 2021 · 45 citations
- Interventional Sum-Product Networks: Causal Inference with Tractable Probabilistic ModelsAuthor 5 of 5NeurIPS 2021 · 42 citations
- Right for the Right Concept: Revising Neuro-Symbolic Concepts by Interacting With Their ExplanationsAuthor 3 of 3CVPR 2021
- Padé Activation Units: End-to-end Learning of Flexible Activation Functions in Deep NetworksAuthor 3 of 3ICLR 2020 · 116 citations
- Structured Object-Aware Physics Prediction for Video Modeling and PlanningAuthor 5 of 5ICLR 2020 · 77 citations
