Peter Súkeník

15 papers A* 6Journal 9

Year	Rank	Type	Title / Venue / Authors
2025	J	`jnl`	Generalization in multi-objective machine learning. Neural Comput. Appl. Peter Súkeník, Christoph Lampert
2025	J	`jnl`	Neural Collapse is Globally Optimal in Deep Regularized ResNets and Transformers. CoRR Peter Súkeník, Christoph H. Lampert, Marco Mondelli
2025	A*	`conf`	Wide Neural Networks Trained with Weight Decay Provably Exhibit Neural Collapse. ICLR Arthur Jacot, Peter Súkeník, Zihan Wang, Marco Mondelli
2024	A*	`conf`	Average gradient outer product as a mechanism for deep neural collapse. NeurIPS Daniel Beaglehole, Peter Súkeník, Marco Mondelli, Mikhail Belkin
2024	J	`jnl`	Average gradient outer product as a mechanism for deep neural collapse. CoRR Daniel Beaglehole, Peter Súkeník, Marco Mondelli, Mikhail Belkin
2024	J	`jnl`	Neural Collapse versus Low-rank Bias: Is Deep Neural Collapse Really Optimal? CoRR Peter Súkeník, Marco Mondelli, Christoph Lampert
2024	A*	`conf`	Neural collapse vs. low-rank bias: Is deep neural collapse really optimal? NeurIPS Peter Súkeník, Christoph H. Lampert, Marco Mondelli
2024	J	`jnl`	Wide Neural Networks Trained with Weight Decay Provably Exhibit Neural Collapse. CoRR Arthur Jacot, Peter Súkeník, Zihan Wang, Marco Mondelli
2023	A*	`conf`	Deep Neural Collapse Is Provably Optimal for the Deep Unconstrained Features Model. NeurIPS Peter Súkeník, Marco Mondelli, Christoph H. Lampert
2023	J	`jnl`	Deep Neural Collapse Is Provably Optimal for the Deep Unconstrained Features Model. CoRR Peter Súkeník, Marco Mondelli, Christoph Lampert
2022	J	`jnl`	Generalization In Multi-Objective Machine Learning. CoRR Peter Súkeník, Christoph H. Lampert
2022	A*	`conf`	Intriguing Properties of Input-Dependent Randomized Smoothing. ICML Peter Súkeník, Aleksei Kuvshinov, Stephan Günnemann
2022	A*	`conf`	The Unreasonable Effectiveness of Fully-Connected Layers for Low-Data Regimes. NeurIPS Peter Kocsis, Peter Súkeník, Guillem Brasó, Matthias Nießner, Laura Leal-Taixé, Ismail Elezi
2022	J	`jnl`	The Unreasonable Effectiveness of Fully-Connected Layers for Low-Data Regimes. CoRR Peter Kocsis, Peter Súkeník, Guillem Brasó, Matthias Nießner, Laura Leal-Taixé, Ismail Elezi
2021	J	`jnl`	Intriguing Properties of Input-dependent Randomized Smoothing. CoRR Peter Súkeník, Aleksei Kuvshinov, Stephan Günnemann