Bernd Rosenow

14 papers A* 2Journal 11Unranked 1

Year	Rank	Type	Title / Venue / Authors
2025	A*	`conf`	Analyzing Neural Scaling Laws in Two-Layer Networks with Power-Law Data Spectra. ICLR Roman Worschech, Bernd Rosenow
2025	J	`jnl`	Anti-correlated noise in epoch-based stochastic gradient descent: implications for weight variances in flat directions. Mach. Learn. Sci. Technol. Marcel Kühn, Bernd Rosenow
2025	A*	`conf`	Enhancing Noise-Robust Losses for Large-Scale Noisy Data Learning. AAAI Max Staats, Matthias Thamm, Bernd Rosenow
2024	J	`jnl`	Analyzing Neural Scaling Laws in Two-Layer Networks with Power-Law Data Spectra. CoRR Roman Worschech, Bernd Rosenow
2024	J	`jnl`	Locating Information in Large Language Models via Random Matrix Theory. CoRR Max Staats, Matthias Thamm, Bernd Rosenow
2023	J	`jnl`	Correlated Noise in Epoch-Based Stochastic Gradient Descent: Implications for Weight Variances. CoRR Marcel Kühn, Bernd Rosenow
2023	J	`jnl`	Online Learning for the Random Feature Model in the Student-Teacher Framework. CoRR Roman Worschech, Bernd Rosenow
2023	J	`jnl`	Reevaluating Loss Functions: Enhancing Robustness to Label Noise in Deep Learning Models. CoRR Max Staats, Matthias Thamm, Bernd Rosenow
2023	J	`jnl`	Topological gap protocol based machine learning optimization of Majorana hybrid wires. CoRR Matthias Thamm, Bernd Rosenow
2022	J	`jnl`	Boundary between noise and information applied to filtering neural network weight matrices. CoRR Max Staats, Matthias Thamm, Bernd Rosenow
2022	J	`jnl`	Machine learning optimization of Majorana hybrid nanowires. CoRR Matthias Thamm, Bernd Rosenow
2022	J	`jnl`	Random matrix analysis of deep neural network weight matrices. CoRR Matthias Thamm, Max Staats, Bernd Rosenow
2021	J	`jnl`	Soft Mode in the Dynamics of Over-realizable On-line Learning for Soft Committee Machines. CoRR Frederieke Richert, Roman Worschech, Bernd Rosenow
2004	—	`conf`	Smooth Correlation Estimation with Application to Portfolio Credit Risk. GfKl Rafael Weißbach, Bernd Rosenow