Index Catalog // ScholarsArchive@OSU

Start Over

Filtering by: Creator Meigs, Andrew J. Committee Member Wong, Weng-Keen Contributor Fu, Xiao

Toward Disentangling the Activations of the Deep Networks via Low-dimensional Embedding and Non-negative Factorization

Creator:

Khorram, Saeed

Abstract:

In this thesis, we introduce a novel Explanation Neural Network (XNN) to explain the predictions made by a deep network. The XNN works by embedding a high-dimensional activation vector of a deep network layer non-linearly into a low-dimensional explanation space while retaining faithfulness i.e., the original deep learning predictions can...

Resource Type:

Masters Thesis

Full Text:

M ,S 1 2 ‖X −MS‖2F (2.8) subject to Mij ≥ 0, Sij ≥ 0 ∀i, j where r is the decomposition rank and