29 citations · 106 across the 17 of their papers we have counts for
3 papers · 1 filter
The Shaped Transformer: Attention Models in the Infinite Depth-and-Width Limit
Lorenzo Noci, Chuning Li, Mufan Bill Li +4
In deep learning theory, the covariance matrix of the representations serves as a proxy to examine the network's trainability. Motivated by the success of Transformers, we study th…
Phenomenology of Double Descent in Finite-Width Neural Networks
Sidak Pal Singh, Aurelien Lucchi, Thomas Hofmann +1
`Double descent' delineates the generalization behaviour of models depending on the regime they belong to: under- or over-parameterized. The current theoretical understanding behin…
Batch Normalization Provably Avoids Rank Collapse for Randomly Initialised Deep Networks
Hadi Daneshmand, Jonas Kohler, Francis Bach +2
Randomly initialized neural networks are known to become harder to train with increasing depth, unless architectural enhancements like residual connections and batch normalization…