AarXiv28 Oct 2015Cited 214×

Deep Learning of Part-Based Representation of Data Using Sparse Autoencoders With Nonnegativity Constraints

Ehsan Hosseini-Asl, Jacek M. Żurada, Olfa Nasraoui

Abstract

1 ©2015 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

A figure from Deep Learning of Part-Based Representation of Data Using Sparse Autoencoders With Nonnegativity Constraints — fig. from the paper

§ The Valyu brief

The paper, in five minutes

Read from the full text via Valyu Contents

A nonnegativity-constrained autoencoder learns part-based representations, improving sparsity and reconstruction quality.

What’s new

i.Introduced a nonnegativity constraint algorithm (NCAE) for deep autoencoders
ii.Demonstrated that NCAE learns part‑based representations of data
iii.Showed improved sparsity and reconstruction quality versus traditional sparse autoencoders and Nonnegative Matrix Factorization
iv.Improved prediction performance of a deep neural network using the learned representation

How they did it

The autoencoder network was trained using a nonnegativity constraint algorithm that forces weights to be non‑negative, effectively limiting negative connections. Performance was evaluated on three standard image datasets and one text dataset to assess part‑based decomposition and prediction accuracy.

What they found

Nonnegativity constraint forces the autoencoder to learn part‑based features and improves sparsity and reconstruction quality compared to traditional sparse autoencoders and NMF

§ The Valyu brief

Reading the full paper and taking notes. This takes a few seconds…

§ Ask this paper

Ask a question about this paper

Valyu reads the full text and answers from what the paper actually says.

§ Related work in other archives

Deep Learning of Part-Based Representation of Data Using Sparse Autoencoders With Nonnegativity Constraints

The paper, in five minutes

Ask a question about this paper

Similar papers in all four archives

Similar papers in all four archives

The paper, in five minutes