Digitized Honors Theses (2002-2017)

Date of Award

5-2015

Document Type

Undergraduate Thesis

Degree Name

BS

Department

Computer Science

Faculty Mentor

Michael Doran, Ph.D.

Advisor(s)

Tom Johnsten, Ph.D., Todd McDonald, Ph.D.

Abstract

We propose a Multimodal Stacked Denoising Autoencoder for learning a joint model of data that consists of multiple modalities. The model is used to extract a joint representation that fuses modalities together. We will evaluate our model on multimodal image classification experiments. Our model is made up of layers of denoising autoencoders which are trained locally to denoise corrupted versions of their inputs. This allows the model to learn useful features from unlabeled data in an unsupervised fashion. We will also investigate generating tags from an image using denoising autoencoders.

Comments

© 2015 Patrick Poirson ALL RIGHTS RESERVED

Download

Share

COinS