ReLU code space: A basis for rating network quality besides accuracy

Autoren Natalia Shepeleva
Werner Zellinger
Michal Mariusz Lewandowski
Bernhard A. Moser
Editoren
Titel ReLU code space: A basis for rating network quality besides accuracy
Buchtitel Machine Learning (strat.ML), https://arxiv.org/abs/2005.09903
Typ in Konferenzband
Organisation Published as a conference paper at NAS 2020 - 1st ICLR Workshop on Neural Architecture Search, Addis Ababa, Ethiopia, April 26, 2020 collocated with ICLR 2020 - 8th International Conference on Learning Representations, Addis Ababa, Ethiopia, April 26-30,
Monat April
Jahr 2020
SCCH ID# 20012
Abstract

We propose a new metric space of ReLU activation codes equipped with a truncated Hamming distance which establishes an isometry between its elements and polyhedral bodies in the input space which have recently been shown to be strongly related to safety, robustness, and confidence. This isometry allows the efficient computation of adjacency relations between the polyhedral bodies. Experiments on MNIST and CIFAR-10 indicate that information besides accuracy might be stored in the code space.