Finite-width executions of Netsor tensor programs converge to their infinite-width Gaussian-process limits in Wasserstein distance at rate O(1/√n) per hidden width, covering weight-sharing architectures including RNNs and attention.
Wide neural networks of any depth evolve as linear models under gradient descent * , volume=
1 Pith paper cite this work, alongside 283 external citations. Polarity classification is still indexing.
1
Pith paper citing it
283
external citations · OpenAlex
fields
cs.LG 1years
2026 1verdicts
ACCEPT 1representative citing papers
citing papers explorer
-
Quantitative Gaussian-Process limits of Tensor Programs
Finite-width executions of Netsor tensor programs converge to their infinite-width Gaussian-process limits in Wasserstein distance at rate O(1/√n) per hidden width, covering weight-sharing architectures including RNNs and attention.