A cosine similarity metric on singular vectors is proposed to evaluate random matrix theory thresholds for singular value truncation in DNN weight matrices.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Evaluating Singular Value Thresholds for DNN Weight Matrices based on Random Matrix Theory
A cosine similarity metric on singular vectors is proposed to evaluate random matrix theory thresholds for singular value truncation in DNN weight matrices.