Saddle-to-saddle dynamics in diagonal linear networks von Pesme, Scott, Flammarion, Nicolas

Volltext

Saddle-to-Saddle Dynamics in Diagonal Linear Networks von Pesme, Scott, Flammarion, Nicolas

Volltext bestellen

Leveraging Continuous Time to Understand Momentum When Training Diagonal Linear Networks von Papazov, Hristo, Pesme, Scott, Flammarion, Nicolas

Volltext bestellen

Saddle-to-Saddle Dynamics in Diagonal Linear Networks von Pesme, Scott, Flammarion, Nicolas

Volltext

(S)GD over Diagonal Linear Networks: Implicit Regularisation, Large Stepsizes and Edge of Stability von Even, Mathieu, Pesme, Scott, Gunasekar, Suriya, Flammarion, Nicolas

Volltext bestellen

Implicit Bias of SGD for Diagonal Linear Networks: a Provable Benefit of Stochasticity von Pesme, Scott, Pillaud-Vivien, Loucas, Flammarion, Nicolas

Volltext bestellen

Leveraging Continuous Time to Understand Momentum When Training Diagonal Linear Networks von Papazov, Hristo, Pesme, Scott, Flammarion, Nicolas

Volltext

Online Robust Regression via SGD on the l1 loss von Pesme, Scott, Flammarion, Nicolas

Volltext

(S)GD over Diagonal Linear Networks: Implicit Regularisation, Large Stepsizes and Edge of Stability von Even, Mathieu, Pesme, Scott, Gunasekar, Suriya, Flammarion, Nicolas

Volltext

On Convergence-Diagnostic based Step Sizes for Stochastic Gradient Descent von Pesme, Scott, Dieuleveut, Aymeric, Flammarion, Nicolas

Volltext

Implicit Bias of SGD for Diagonal Linear Networks: a Provable Benefit of Stochasticity von Pesme, Scott, Pillaud-Vivien, Loucas, Flammarion, Nicolas

Volltext bestellen