bwbellmath

2 ms·

1. 1.▲
   An Exact Equivalence for Finite Classification Models(arxiv.org)

   2 points by bwbellmath 3y ago|1 comments
2. 2.▲
   by bwbellmath 3y ago

   We re-explore the path kernel result from Domingos and address some errors and limitations from his approach to derive an exact and practical kernel representation.
3. 3.▲
   by bwbellmath 3y ago

   See our re-examination of the kernel equivalence. Path kernels exactly measure how models learn as their understanding of data improves during training, and this can be expressed in terms of the gradients with regards to each trianing input