Accelerating Convergence of Stein Variational Gradient Descent via Deep Unfoldingby Yuya Kawamura, Satoshi TakabeFirst submitted…
Machine Unlearning of Pre-trained Large Language Modelsby Jin Yao, Eli Chien, Minxin Du, Xinyao Niu,…
In-Context Learning of a Linear Transformer Block: Benefits of the MLP Component and One-Step GD…
Energy-efficiency Limits on Training AI Systems using Learning-in-Memoryby Zihao Chen, Johannes Leugering, Gert Cauwenberghs, Shantanu…
Linear Transformers are Versatile In-Context Learnersby Max Vladymyrov, Johannes von Oswald, Mark Sandler, Rong GeFirst…
Private Gradient Descent for Linear Regression: Tighter Error Bounds and Instance-Specific Uncertainty Estimationby Gavin Brown,…
SDEs for Minimax Optimizationby Enea Monzio Compagnoni, Antonio Orvieto, Hans Kersting, Frank Norbert Proske, Aurelien…
Convergence of Gradient Descent for Recurrent Neural Networks: A Nonasymptotic Analysisby Semih Cayci, Atilla EryilmazFirst…
Stochastic Hessian Fittings with Lie Groupsby Xi-Lin LiFirst submitted to arxiv on: 19 Feb 2024CategoriesMain:…
LoRA Training in the NTK Regime has No Spurious Local Minimaby Uijeong Jang, Jason D.…