| Deep neural collapse is provably optimal for the deep unconstrained features model P Súkeník, M Mondelli, CH Lampert Advances in Neural Information Processing Systems 36, 52991-53024, 2023 | 36 | 2023 |
| Intriguing Properties of Input-dependent Randomized Smoothing P Súkeník, A Kuvshinov, S Günnemann Proceedings of the 39th International Conference on Machine Learning, PMLR …, 2021 | 36 | 2021 |
| Average gradient outer product as a mechanism for deep neural collapse D Beaglehole, P Súkeník, M Mondelli, M Belkin Advances in Neural Information Processing Systems 37, 130764-130796, 2024 | 21 | 2024 |
| Neural collapse vs. low-rank bias: Is deep neural collapse really optimal? P Súkeník, C Lampert, M Mondelli Advances in Neural Information Processing Systems 37, 138250-138288, 2024 | 19 | 2024 |
| Generalization in multi-objective machine learning P Súkeník, C Lampert Neural Computing and Applications, 1-15, 2024 | 13 | 2024 |
| The Unreasonable Effectiveness of Fully-Connected Layers for Low-Data Regimes P Kocsis, P Súkeník, G Brasó, M Nießner, L Leal-Taixé, I Elezi Advances in Neural Information Processing Systems 35 (NeurIPS 2022), 2022 | 10 | 2022 |
| Wide neural networks trained with weight decay provably exhibit neural collapse A Jacot, P Súkeník, Z Wang, M Mondelli arXiv preprint arXiv:2410.04887, 2024 | 8 | 2024 |
| Neural Collapse is Globally Optimal in Deep Regularized ResNets and Transformers P Súkeník, CH Lampert, M Mondelli arXiv preprint arXiv:2505.15239, 2025 | 1 | 2025 |
Marco MondelliProfessor, IST AustriaVerified email at ist.ac.at
