Matthias Hein 0001
90 publications
14 venues
H Index 40
Affiliation
Max Planck Institute for Intelligent Systems, T bingen, GermanyFaculty of Mathematics and Computer Science, Saarland University
Max Planck Institute for Biological Cybernetics, T bingen, Germany
Links
| Name | Venue | Year | citations |
|---|---|---|---|
Mahalanobis++: Improving OOD Detection via Feature Normalization.
|
ICML | 2025 | 9 |
| An Interpretable N-gram Perplexity Threat Model for Large Language Model Jailbreaks.
|
ICML | 2025 | 0 |
| Bias of Stochastic Gradient Descent or the Architecture: Disentangling the Effects of Overparameterization of Neural Networks.
|
ICML | 2024 | 0 |
| Robust CLIP: Unsupervised Adversarial Fine-Tuning of Vision Embeddings for Robust Large Vision-Language Models.
|
ICML | 2024 | 96 |
| Towards Reliable Evaluation and Fast Training of Robust Semantic Segmentation Models.
|
ECCV | 2024 | 0 |
| DiG-IN: Diffusion Guidance for Investigating Networks - Uncovering Classifier Differences, Neuron Visualisations, and Visual Counterfactual Explanations.
|
CVPR | 2024 | 0 |
| In or Out? Fixing ImageNet Out-of-Distribution Detection Evaluation.
|
ICML | 2023 | 137 |
| A Modern Look at the Relationship between Sharpness and Generalization.
|
ICML | 2023 | 87 |
| Revisiting Adversarial Training for ImageNet: Architectures, Training and Generalization across Threat Models.
|
NIPS/NeurIPS | 2023 | 99 |
| Improving l1-Certified Robustness via Randomized Smoothing by Leveraging Box Constraints.
|
ICML | 2023 | 11 |
| Normalization Layers Are All That Sharpness-Aware Minimization Needs.
|
NIPS/NeurIPS | 2023 | 35 |
| Sound Randomized Smoothing in Floating-Point Arithmetic.
|
ICLR | 2023 | 0 |
| Certified Defences Against Adversarial Patch Attacks on Semantic Segmentation.
|
ICLR | 2023 | 0 |
| Spurious Features Everywhere - Large-Scale Detection of Harmful Spurious Features in ImageNet.
|
ICCV | 2023 | 0 |
| Random and Adversarial Bit Error Robustness: Energy-Efficient and Secure DNN Accelerators.
|
TPAMI | 2023 | 0 |
| Evaluating the Adversarial Robustness of Adaptive Test-time Defenses.
|
ICML | 2022 | 81 |
| Breaking Down Out-of-Distribution Detection: Many Methods Based on OOD Training Data Estimate a Combination of the Same Core Quantities.
|
ICML | 2022 | 35 |
| Provably Adversarially Robust Nearest Prototype Classifiers.
|
ICML | 2022 | 15 |
| Diffusion Visual Counterfactual Explanations.
|
NIPS/NeurIPS | 2022 | 104 |
| Neural Network Heuristic Functions: Taking Confidence into Account.
|
SOCS | 2022 | 6 |
| Sparse-RS: A Versatile Framework for Query-Efficient Sparse Black-Box Adversarial Attacks.
|
AAAI | 2022 | 0 |
| Adversarial Robustness against Multiple and Single l
|
ICML | 2022 | 0 |
| Being a Bit Frequentist Improves Bayesian Neural Networks.
|
AISTATS | 2022 | 0 |
| Provably Adversarially Robust Detection of Out-of-Distribution Data (Almost) for Free.
|
NIPS/NeurIPS | 2022 | 0 |
| Meta-Learning the Search Distribution of Black-Box Random Search Based Adversarial Attacks.
|
NIPS/NeurIPS | 2021 | 15 |
| Relating Adversarially Robust Generalization to Flat Minima.
|
ICCV | 2021 | 78 |
| Mind the Box: l
|
ICML | 2021 | 0 |
| Learnable uncertainty under Laplace approximations.
|
UAI | 2021 | 0 |
| An Infinite-Feature Extension for Bayesian ReLU Nets That Fixes Their Asymptotic Overconfidence.
|
NIPS/NeurIPS | 2021 | 0 |
| Being Bayesian, Even Just a Bit, Fixes Overconfidence in ReLU Networks.
|
ICML | 2020 | 0 |
| Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attacks.
|
ICML | 2020 | 2260 |
| Certifiably Adversarially Robust Detection of Out-of-Distribution Data.
|
NIPS/NeurIPS | 2020 | 83 |
| Adversarial Robustness on In- and Out-Distribution Improves Explainability.
|
ECCV | 2020 | 109 |
| Minimally distorted Adversarial Examples with a Fast Adaptive Boundary Attack.
|
ICML | 2020 | 0 |
| Confidence-Calibrated Adversarial Training: Generalizing to Unseen Attacks.
|
ICML | 2020 | 0 |
| Square Attack: A Query-Efficient Black-Box Adversarial Attack via Random Search.
|
ECCV | 2020 | 0 |
| Towards neural networks that provably know when they don't know.
|
ICLR | 2020 | 0 |
| Provable robustness against all adversarial $l_p$-perturbations for $p\geq 1$.
|
ICLR | 2020 | 0 |
| Provably robust boosted decision stumps and trees against adversarial attacks.
|
NIPS/NeurIPS | 2019 | 68 |
| Sparse and Imperceivable Adversarial Attacks.
|
ICCV | 2019 | 223 |
| Generalized Matrix Means for Semi-Supervised Learning with Multilayer Graphs.
|
NIPS/NeurIPS | 2019 | 19 |
| Spectral Clustering of Signed Graphs via Matrix Power Means.
|
ICML | 2019 | 37 |
| Provable Robustness of ReLU networks via Maximization of Linear Regions.
|
AISTATS | 2019 | 0 |
| Disentangling Adversarial Robustness and Generalization.
|
CVPR | 2019 | 0 |
| Why ReLU Networks Yield High-Confidence Predictions Far Away From the Training Data and How to Mitigate the Problem.
|
CVPR | 2019 | 0 |
| The Power Mean Laplacian for Multilayer Graph Clustering.
|
AISTATS | 2018 | 30 |
| Neural Networks Should Be Wide Enough to Learn Disconnected Decision Regions.
|
ICML | 2018 | 57 |
| Optimization Landscape and Expressivity of Deep CNNs.
|
ICML | 2018 | 0 |
| Analysis and Optimization of Loss Functions for Multiclass, Top-k, and Multilabel Classification.
|
TPAMI | 2018 | 0 |
| Variants of RMSProp and Adagrad with Logarithmic Regret Bounds.
|
ICML | 2017 | 270 |
| The Loss Surface of Deep and Wide Neural Networks.
|
ICML | 2017 | 297 |
| Formal Guarantees on the Robustness of a Classifier against Adversarial Manipulation.
|
NIPS/NeurIPS | 2017 | 533 |
| Simple Does It: Weakly Supervised Instance and Semantic Segmentation.
|
CVPR | 2017 | 0 |
| An Efficient Multilinear Optimization Framework for Hypergraph Matching.
|
TPAMI | 2017 | 0 |
| Globally Optimal Training of Generalized Polynomial Neural Networks with Nonlinear Spectral Methods.
|
NIPS/NeurIPS | 2016 | 33 |
| Clustering Signed Networks with the Geometric Mean of Laplacians.
|
NIPS/NeurIPS | 2016 | 50 |
| Latent Embeddings for Zero-Shot Classification.
|
CVPR | 2016 | 745 |
| Weakly Supervised Object Boundaries.
|
CVPR | 2016 | 0 |
| Loss Functions for Top-k Error: Analysis and Insights.
|
CVPR | 2016 | 0 |
| Regularization-Free Estimation in Trace Regression with Symmetric Positive Semidefinite Matrices.
|
NIPS/NeurIPS | 2015 | 16 |
| Efficient Output Kernel Learning for Multiple Tasks.
|
NIPS/NeurIPS | 2015 | 36 |
| Classifier based graph construction for video segmentation.
|
CVPR | 2015 | 66 |
| A flexible tensor block coordinate ascent scheme for hypergraph matching.
|
CVPR | 2015 | 69 |
| Top-k Multiclass SVM.
|
NIPS/NeurIPS | 2015 | 97 |
| Hitting and commute times in large random neighborhood graphs.
|
JMLR | 2014 | 119 |
| Scalable Multitask Representation Learning for Scene Classification.
|
CVPR | 2014 | 57 |
| Tight Continuous Relaxation of the Balanced k-Cut Problem.
|
NIPS/NeurIPS | 2014 | 19 |
| The Total Variation on Hypergraphs - Learning on Hypergraphs Revisited.
|
NIPS/NeurIPS | 2013 | 168 |
| Towards realistic team formation in social networks based on densest subgraphs.
|
WWW | 2013 | 99 |
| Constrained fractional set programs and their application in local clustering and community detection.
|
ICML | 2013 | 21 |
| Matrix factorization with binary components.
|
NIPS/NeurIPS | 2013 | 41 |
| Constrained 1-Spectral Clustering.
|
AISTATS | 2012 | 88 |
| Beyond Spectral Clustering - Tight Relaxations of Balanced Graph Cuts.
|
NIPS/NeurIPS | 2011 | 92 |
| Sparse recovery by thresholded non-negative least squares.
|
NIPS/NeurIPS | 2011 | 0 |
| An Inverse Power Method for Nonlinear Eigenproblems with Applications in 1-Spectral Clustering and Sparse PCA.
|
NIPS/NeurIPS | 2010 | 220 |
| Getting lost in space: Large sample analysis of the resistance distance.
|
NIPS/NeurIPS | 2010 | 79 |
| Semi-supervised Regression using Hessian energy with an application to semi-supervised dimensionality reduction.
|
NIPS/NeurIPS | 2009 | 113 |
| Robust Nonparametric Regression with Metric-Space Valued Output.
|
NIPS/NeurIPS | 2009 | 42 |
| Spectral clustering based on the graph
|
ICML | 2009 | 0 |
| Non-parametric Regression Between Manifolds.
|
NIPS/NeurIPS | 2008 | 62 |
| Influence of graph construction on graph-based clustering measures.
|
NIPS/NeurIPS | 2008 | 186 |
| Manifold Denoising as Preprocessing for Finding Natural Representations of Data.
|
AAAI | 2007 | 16 |
| Graph Laplacians and their Convergence on Random Neighborhood Graphs.
|
JMLR | 2007 | 0 |
| Manifold Denoising.
|
NIPS/NeurIPS | 2006 | 207 |
| Uniform Convergence of Adaptive Graph-Based Regularization.
|
COLT | 2006 | 84 |
| Hilbertian Metrics and Positive Definite Kernels on Probability Measures.
|
AISTATS | 2005 | 230 |
| Intrinsic dimensionality estimation of submanifolds in R
|
ICML | 2005 | 73 |
| From Graphs to Manifolds - Weak and Strong Pointwise Consistency of Graph Laplacians.
|
COLT | 2005 | 363 |
| Measure Based Regularization.
|
NIPS/NeurIPS | 2003 | 140 |
| Maximal Margin Classification for Metric Spaces.
|
COLT | 2003 | 0 |