Loss Landscape
PulseAugur coverage of Loss Landscape — every cluster mentioning Loss Landscape across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
Neural network dissertation probes loss landscape and feature learning
A dissertation explores the internal workings of neural networks, focusing on their optimization processes and the phenomenon of mode connectivity within the loss landscape. The research aims to demystify how these netw…
-
New diagnostic tool assesses symmetry learning in neural PDE emulators
Researchers have developed a new diagnostic tool to assess how well neural emulators of partial differential equations internalize physical symmetries. This method measures the propagation of parameter updates between s…
-
New DSCH-Loss method enhances deep semantic hashing performance
Researchers have introduced DSCH-Loss, a novel objective function for deep semantic hashing that aims to improve the efficiency and accuracy of approximate nearest neighbor search. Unlike previous methods that used fixe…
-
New 'weight-norm criticality' explains AI training instability
Researchers have identified a new critical factor in deep neural network training instability, termed 'weight-norm criticality.' This phenomenon, distinct from the commonly understood 'learning-rate criticality,' arises…
-
Hessian Spectrum of Neural Networks Tied to Data Distribution
A new research paper published on arXiv explores the relationship between the Hessian matrix's spectrum and the data used in deep learning models. The study derives eigenvalues for linear networks, revealing that for cl…
-
New research frameworks model gradient descent at the edge of stability
Two new research papers explore the phenomenon of gradient descent operating at the edge of stability (EoS) in deep learning. The first paper introduces 'Edge Flow,' a system of differential equations that models gradie…