Numerical Analysis, Floating Point, Precision Guarantees, Computational Mathematics
On the computation of tensor functions under tensor-tensor multiplications with linear maps
arxiv.org·3d
Tight Generalization Error Bounds for Stochastic Gradient Descent in Non-convex Learning
arxiv.org·3d
EQuARX: Efficient Quantized AllReduce in XLA for Distributed Machine Learning Acceleration
arxiv.org·3d
Discrete-Time Linear Dynamical System Control Using Sparse Inputs With Time-Varying Support
arxiv.org·3d
Loading...Loading more...