BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Date iCal//NONSGML kigkonsult.se iCalcreator 2.20.4//
METHOD:PUBLISH
X-WR-CALNAME;VALUE=TEXT:ÌÇÐÄÔ­´´
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:STANDARD
DTSTART:20191103T020000
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20200308T020000
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:calendar.377796.field_event_date.0@www.wright.edu
DTSTAMP:20260220T000916Z
CREATED:20191113T170606Z
DESCRIPTION:Ph.D. Committee:  Drs. Derek Doran (advisor)\, Pascal Hitzler (
 Kansas State University)\, Michael Raymer\, and Tanvi Banerjee ABSTRACT:De
 ep Neural Networks (DNNs) are powerful tools blossomed in a variety of suc
 cessful real-life applications. While the performance of DNNs is outstandi
 ng\, their opaque nature raises a growing concern in the community\, causi
 ng suspicions on the reliability and trustworthiness of decisions made by 
 DNNs. In order to release such concerns and towards building reliable deep
  learning systems\, research efforts are actively made in diverse aspects 
 such as model interpretation\, model fairness and bias\, adversarial attac
 ks and defenses\, and so on.In this dissertation\, we focus on the researc
 h topic of DNN interpretations\, aiming to unfold the black-box and provid
 e explanations in a human-understandable way. We first conduct a categoriz
 ed literature review\, introducing the realm of explainable deep learning.
  Following the review\, two specific problems are tackled\, explanations o
 f Convolutions Neural Networks (CNNs)\, which relates the CNN decisions wi
 th input concepts\, and interpretability of multi-model interactions\, whe
 re an explainable model is built to solve a task similar to visual questio
 n answering. Visualization techniques are leveraged to depict the intermed
 iate hidden states of CNNs and attention mechanisms are utilized to build 
 an instinct explainable model. Towards increasing the trustworthiness of D
 NNs\, a certainty measurement for decisions are also proposed as a potenti
 al future extension of this study.
DTSTART;TZID=America/New_York:20191120T140000
DTEND;TZID=America/New_York:20191120T160000
LAST-MODIFIED:20191118T125910Z
LOCATION:399 Joshi
SUMMARY:Ph.D. Dissertation Proposal Defense Towards Interpretable and Relia
 ble Deep Neural Networks By Ning Xie
URL;TYPE=URI:/events/phd-dissertation-proposal-defens
 e-towards-interpretable-reliable-deep-neural-networks-ning
END:VEVENT
END:VCALENDAR