ASMS_CSC Archives

MS Student

ASMS_CSC@LISTSERV.GSU.EDU

Options: Use Classic View

Use Monospaced Font
Show Text Part by Default
Condense Mail Headers

Topic: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Sender: MS Student <[log in to unmask]>
Date: Mon, 9 May 2022 20:49:47 +0000
Reply-To: MS Student <[log in to unmask]>
Message-ID: <[log in to unmask]>
From: Jamie Hayes <[log in to unmask]>
MIME-Version: 1.0
In-Reply-To: <[log in to unmask]>
Content-Type: multipart/mixed; boundary="_003_BN7PR05MB6274E82890C1247728D21B0CC8C69BN7PR05MB6274namp_"
Parts/Attachments: text/plain (1220 bytes) , NTT Interview Wang Peng.docx (16 kB) , Wang_Peng_portPDF_4456190.pdf (98 kB)
Dear Graduate Students´╝î

    Enclosed please find the schedule and CV for our NTT candidate, Peng Wang on May 10 (Tomorrow). The talk includes two parts 1) 30 minutes research talk, 2) 30 minutes lecture on recursion.

Title: Threat and Protection on Image-Based Deep Learning

Abstract: In recent years, a series of researches have revealed that Deep Neural Network (DNN) is vulnerable to adversarial attack, and a number of attack methods have been proposed. Among those methods, an extremely sly type of attack named one-pixel attack can mislead DNNs to misclassify an image via only modifying one pixel of the image, leading to severe security threats to DNN-based information systems.
My work reveals the internal threat of one-pixel attack and proposed two detection methods, including trigger detection and candidate detection, that filled the blank of no method can really detect one-pixel attack.
The trigger detection method analyzes the vulnerability of DNN models and gives the most suspected pixel that is modified by one-pixel attack.
The candidate detection method identifies a set of most suspected pixels using differential evolution-based heuristic algorithm.

Best Regards

Zhipeng


ATOM RSS1 RSS2