Felix Dangel - RankMe

38 papers A* 11A 1Journal 25

Year	Rank	Type	Title / Venue / Authors
2026	J	`jnl`	Dataless Weight Disentanglement in Task Arithmetic via Kronecker-Factored Approximate Curvature. CoRR Angelo Porrello, Pietro Buzzega, Felix Dangel, Thomas Sommariva, Riccardo Salami, Lorenzo Bonicelli, Simone Calderara
2026	J	`jnl`	Efficient Bilevel Optimization with KFAC-Based Hypergradients. CoRR Disen Liao, Felix Dangel, Yaoliang Yu
2025	J	`jnl`	Collapsing Taylor Mode Automatic Differentiation. CoRR Felix Dangel, Tim Siebert, Marius Zeinhofer, Andrea Walther
2025	A*	`conf`	Fishers for Free? Approximating the Fisher Information Matrix by Recycling the Squared Gradient Accumulator. ICML Yu Xin Li, Felix Dangel, Derek Tam, Colin Raffel
2025	J	`jnl`	Fishers for Free? Approximating the Fisher Information Matrix by Recycling the Squared Gradient Accumulator. CoRR Yuxin Li, Felix Dangel, Derek Tam, Colin Raffel
2025	A*	`conf`	Hide & Seek: Transformer Symmetries Obscure Sharpness & Riemannian Geometry Finds It. ICML Marvin F. da Silva, Felix Dangel, Sageev Oore
2025	J	`jnl`	Hide & Seek: Transformer Symmetries Obscure Sharpness & Riemannian Geometry Finds It. CoRR Marvin F. da Silva, Felix Dangel, Sageev Oore
2025	J	`jnl`	Improving Energy Natural Gradient Descent through Woodbury, Momentum, and Randomization. CoRR Andrés Guzmán-Cordero, Felix Dangel, Gil Goldshlager, Marius Zeinhofer
2025	J	`jnl`	Kronecker-factored Approximate Curvature (KFAC) From Scratch. CoRR Felix Dangel, Bálint Mucsányi, Tobias Weber, Runa Eschenhagen
2025	J	`jnl`	Position: Curvature Matrices Should Be Democratized via Linear Operators. CoRR Felix Dangel, Runa Eschenhagen, Weronika Ormaniec, Andres Fernandez, Lukas Tatzel, Agustinus Kristiadi
2025	J	`jnl`	Sketching Low-Rank Plus Diagonal Matrices. CoRR Andres Fernandez, Felix Dangel, Philipp Hennig, Frank Schneider
2025	J	`jnl`	Spectral-factorized Positive-definite Curvature Learning for NN Training. CoRR Wu Lin, Felix Dangel, Runa Eschenhagen, Juhan Bae, Richard E. Turner, Roger B. Grosse
2025	J	`jnl`	Understanding and Improving Shampoo and SOAP via Kullback-Leibler Minimization. CoRR Wu Lin, Scott C. Lowe, Felix Dangel, Runa Eschenhagen, Zikun Xu, Roger B. Grosse
2025	A*	`conf`	What Does It Mean to Be a Transformer? Insights from a Theoretical Hessian Analysis. ICLR Weronika Ormaniec, Felix Dangel, Sidak Pal Singh
2024	A*	`conf`	Can We Remove the Square-Root in Adaptive Gradient Methods? A Second-Order Perspective. ICML Wu Lin, Felix Dangel, Runa Eschenhagen, Juhan Bae, Richard E. Turner, Alireza Makhzani
2024	J	`jnl`	Can We Remove the Square-Root in Adaptive Gradient Methods? A Second-Order Perspective. CoRR Wu Lin, Felix Dangel, Runa Eschenhagen, Juhan Bae, Richard E. Turner, Alireza Makhzani
2024	A*	`conf`	Convolutions and More as Einsum: A Tensor Network Perspective with Advances for Second-Order Methods. NeurIPS Felix Dangel
2024	A*	`conf`	Kronecker-Factored Approximate Curvature for Physics-Informed Neural Networks. NeurIPS Felix Dangel, Johannes Müller, Marius Zeinhofer
2024	J	`jnl`	Kronecker-Factored Approximate Curvature for Physics-Informed Neural Networks. CoRR Felix Dangel, Johannes Müller, Marius Zeinhofer
2024	J	`jnl`	Lowering PyTorch's Memory Consumption for Selective Differentiation. CoRR Samarth Bhatia, Felix Dangel
2024	A*	`conf`	Revisiting Scalable Hessian Diagonal Approximations for Applications in Reinforcement Learning. ICML Mohamed Elsayed, Homayoon Farrahi, Felix Dangel, A. Rupam Mahmood
2024	J	`jnl`	Revisiting Scalable Hessian Diagonal Approximations for Applications in Reinforcement Learning. CoRR Mohamed Elsayed, Homayoon Farrahi, Felix Dangel, A. Rupam Mahmood
2024	A*	`conf`	Structured Inverse-Free Natural Gradient Descent: Memory-Efficient & Numerically-Stable KFAC. ICML Wu Lin, Felix Dangel, Runa Eschenhagen, Kirill Neklyudov, Agustinus Kristiadi, Richard E. Turner, Alireza Makhzani
2024	J	`jnl`	What Does It Mean to Be a Transformer? Insights from a Theoretical Hessian Analysis. CoRR Weronika Ormaniec, Felix Dangel, Sidak Pal Singh
2023	—	`—`	Backpropagation Beyond the Gradient. Felix Dangel
2023	J	`jnl`	Convolutions Through the Lens of Tensor Networks. CoRR Felix Dangel
2023	J	`jnl`	On the Disconnect Between Theory and Practice of Overparametrized Neural Networks. CoRR Jonathan Wenger, Felix Dangel, Agustinus Kristiadi
2023	J	`jnl`	Structured Inverse-Free Natural Gradient: Memory-Efficient & Numerically-Stable KFAC for Large Neural Nets. CoRR Wu Lin, Felix Dangel, Runa Eschenhagen, Kirill Neklyudov, Agustinus Kristiadi, Richard E. Turner, Alireza Makhzani
2023	A*	`conf`	The Geometry of Neural Nets' Parameter Spaces Under Reparametrization. NeurIPS Agustinus Kristiadi, Felix Dangel, Philipp Hennig
2023	J	`jnl`	The Geometry of Neural Nets' Parameter Spaces Under Reparametrization. CoRR Agustinus Kristiadi, Felix Dangel, Philipp Hennig
2023	J	`jnl`	ViViT: Curvature Access Through The Generalized Gauss-Newton's Low-Rank Structure. Trans. Mach. Learn. Res. Felix Dangel, Lukas Tatzel, Philipp Hennig
2021	J	`jnl`	Cockpit: A Practical Debugging Tool for Training Deep Neural Networks. CoRR Frank Schneider, Felix Dangel, Philipp Hennig
2021	A*	`conf`	Cockpit: A Practical Debugging Tool for the Training of Deep Neural Networks. NeurIPS Frank Schneider, Felix Dangel, Philipp Hennig
2021	J	`jnl`	ViViT: Curvature access through the generalized Gauss-Newton's low-rank structure. CoRR Felix Dangel, Lukas Tatzel, Philipp Hennig
2020	A*	`conf`	BackPACK: Packing more into Backprop. ICLR Felix Dangel, Frederik Kunstner, Philipp Hennig
2020	A	`conf`	Modular Block-diagonal Curvature Approximations for Feedforward Architectures. AISTATS Felix Dangel, Stefan Harmeling, Philipp Hennig
2019	J	`jnl`	A Modular Approach to Block-diagonal Hessian Approximations for Second-order Optimization Methods. CoRR Felix Dangel, Philipp Hennig
2019	J	`jnl`	BackPACK: Packing more into backprop. CoRR Felix Dangel, Frederik Kunstner, Philipp Hennig