Summary of Geoadaler: Geometric Insights Into Adaptive Stochastic Gradient Descent Algorithms, by Chinedu Eleh et al.
GeoAdaLer: Geometric Insights into Adaptive Stochastic Gradient Descent Algorithmsby Chinedu Eleh, Masuzyo Mwanza, Ekene Aguegboh,…
GeoAdaLer: Geometric Insights into Adaptive Stochastic Gradient Descent Algorithmsby Chinedu Eleh, Masuzyo Mwanza, Ekene Aguegboh,…
4+3 Phases of Compute-Optimal Neural Scaling Lawsby Elliot Paquette, Courtney Paquette, Lechao Xiao, Jeffrey PenningtonFirst…
Banded Square Root Matrix Factorization for Differentially Private Model Trainingby Nikita P. Kalinin, Christoph LampertFirst…
Almost sure convergence rates of stochastic gradient methods under gradient dominationby Simon Weissmann, Sara Klein,…
Score-based Generative Models with Adaptive Momentumby Ziqing Wen, Xiaoge Deng, Ping Luo, Tao Sun, Dongsheng…
Uncertainty quantification by block bootstrap for differentially private stochastic gradient descentby Holger Dette, Carina GrawFirst…
Flattened one-bit stochastic gradient descent: compressed distributed optimization with controlled varianceby Alexander Stollenwerk, Laurent JacquesFirst…
Random Scaling and Momentum for Non-smooth Non-convex Optimizationby Qinzi Zhang, Ashok CutkoskyFirst submitted to arxiv…
Secure Aggregation Meets Sparsification in Decentralized Learningby Sayan Biswas, Anne-Marie Kermarrec, Rafael Pires, Rishi Sharma,…
VALID: a Validated Algorithm for Learning in Decentralized Networks with Possible Adversarial Presenceby Mayank Bakshi,…