BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:Asia/Calcutta
BEGIN:STANDARD
DTSTART:19451014T230000
TZOFFSETFROM:+0630
TZOFFSETTO:+0530
TZNAME:IST
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20191017T062917Z
UID:91266AB3-C54F-4BD8-84D4-D06E5AC0FB05
DTSTART;TZID=Asia/Calcutta:20190628T103000
DTEND;TZID=Asia/Calcutta:20190628T122600
DESCRIPTION:Expert lecture by Dr. Vineet N. Balasubramanian\, Associate Pro
 fessor\, IIT Hyderabad. He gave a very interesting lecture on how to look 
 inside the CNN\nand try to know what CNN is trying to learn. His topic was
  ‘Explainable Deep Learning Machines’. He\, therefore\, discussed vari
 ous clever methods to visualize what CNN has learned during the training s
 ession. He presented methods using gradients and weights of the filters to
  visualize what the CNN is learning. He then presented a topic called Adve
 rsarial Perturbation which is a process of perturbing the pixels of an ima
 ge to fool the CNN while training.\n\nCo-sponsored by: IIIT - ALLAHABAD\n\
 nSpeaker(s): Vineeth N Balasubramanian\, \n\nRoom: 5005\, Bldg: cc3\, Indi
 an Institute of information technology\, Allahabad\, Uttar Pradesh\, India
 \, 211015
LOCATION:Room: 5005\, Bldg: cc3\, Indian Institute of information technolog
 y\, Allahabad\, Uttar Pradesh\, India\, 211015
ORGANIZER:pis2016004@iiita.ac.in
SEQUENCE:2
SUMMARY: Expert Lectures on ‘Explainable Deep Learning Machines’ By Dr.
  Vineeth N Balasubramanian
URL;VALUE=URI:https://events.vtools.ieee.org/m/207435
X-ALT-DESC:Description: <br /><p>Expert lecture by Dr. Vineet N. Balasubram
 anian\, Associate Professor\, IIT Hyderabad. He gave a very interesting le
 cture on how to look inside the CNN<br />and try to know what CNN is tryin
 g to learn. His topic was &lsquo\;Explainable Deep Learning Machines&rsquo
 \;. He\, therefore\, discussed various clever methods to visualize what CN
 N has learned during the training session. He presented methods using grad
 ients and weights of the filters to visualize what the CNN is learning. He
  then presented a topic called Adversarial Perturbation which is a process
  of perturbing the pixels of an image to fool the CNN while training.</p>
END:VEVENT
END:VCALENDAR