On foveation of deep neural networks
This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
Main Author: | |
---|---|
Other Authors: | |
Format: | Thesis |
Language: | eng |
Published: |
Massachusetts Institute of Technology
2019
|
Subjects: | |
Online Access: | https://hdl.handle.net/1721.1/123134 |
_version_ | 1811095511160061952 |
---|---|
author | Srivastava, Sanjana. |
author2 | Tomaso Poggio. |
author_facet | Tomaso Poggio. Srivastava, Sanjana. |
author_sort | Srivastava, Sanjana. |
collection | MIT |
description | This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. |
first_indexed | 2024-09-23T16:17:33Z |
format | Thesis |
id | mit-1721.1/123134 |
institution | Massachusetts Institute of Technology |
language | eng |
last_indexed | 2024-09-23T16:17:33Z |
publishDate | 2019 |
publisher | Massachusetts Institute of Technology |
record_format | dspace |
spelling | mit-1721.1/1231342019-12-06T03:01:47Z On foveation of deep neural networks Srivastava, Sanjana. Tomaso Poggio. Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science. Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Electrical Engineering and Computer Science. This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2019 Cataloged from student-submitted PDF version of thesis. Includes bibliographical references (pages 61-63). The human ability to recognize objects is impaired when the object is not shown in full. "Minimal images" are the smallest regions of an image that remain recognizable for humans. [26] show that a slight modification of the location and size of the visible region of the minimal image produces a sharp drop in human recognition accuracy. In this paper, we demonstrate that such drops in accuracy due to changes of the visible region are a common phenomenon between humans and existing state-of- the-art convolutional neural networks (CNNs), and are much more prominent in CNNs. We found many cases where CNNs classified one region correctly and the other incorrectly, though they only differed by one row or column of pixels, and were often bigger than the average human minimal image size. We show that this phenomenon is independent from previous works that have reported lack of invariance to minor modifications in object location in CNNs. Our results thus reveal a new failure mode of CNNs that also affects humans to a lesser degree. They expose how fragile CNN recognition ability is for natural images even without synthetic adversarial patterns being introduced. This opens potential for CNN robustness in natural images to be brought to the human level by taking inspiration from human robustness methods. One of these is eccentricity dependence, a model of human focus in which attention to the visual input degrades proportional to distance from the focal point [7]. We demonstrate that applying the "inverted pyramid" eccentricity method, a multi-scale input transformation, makes CNNs more robust to useless background features than a standard raw-image input. Our results also find that using the inverted pyramid method generally reduces useless background pixels, therefore reducing required training data. by Sanjana Srivastava. M. Eng. M.Eng. Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science 2019-12-05T18:05:20Z 2019-12-05T18:05:20Z 2019 2019 Thesis https://hdl.handle.net/1721.1/123134 1128816526 eng MIT theses are protected by copyright. They may be viewed, downloaded, or printed from this source but further reproduction or distribution in any format is prohibited without written permission. http://dspace.mit.edu/handle/1721.1/7582 63 pages application/pdf Massachusetts Institute of Technology |
spellingShingle | Electrical Engineering and Computer Science. Srivastava, Sanjana. On foveation of deep neural networks |
title | On foveation of deep neural networks |
title_full | On foveation of deep neural networks |
title_fullStr | On foveation of deep neural networks |
title_full_unstemmed | On foveation of deep neural networks |
title_short | On foveation of deep neural networks |
title_sort | on foveation of deep neural networks |
topic | Electrical Engineering and Computer Science. |
url | https://hdl.handle.net/1721.1/123134 |
work_keys_str_mv | AT srivastavasanjana onfoveationofdeepneuralnetworks |