BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:America/Edmonton
BEGIN:DAYLIGHT
DTSTART:20220313T030000
TZOFFSETFROM:-0700
TZOFFSETTO:-0600
RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3
TZNAME:MDT
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20221106T010000
TZOFFSETFROM:-0600
TZOFFSETTO:-0700
RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11
TZNAME:MST
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20220421T182100Z
UID:F7B10E76-9EBE-425F-9401-BC91AFE6DECD
DTSTART;TZID=America/Edmonton:20220420T150000
DTEND;TZID=America/Edmonton:20220420T160000
DESCRIPTION:Machine learning requires assumptions and data\, with the assum
 ptions based on domain knowledge. With the exponential increase in dataset
  set and computation we’ve been able to reduce the domain expertise and 
 inductive bias we integrate in computer vision models\, from manually desi
 gned features such as SIFT\, to learning representations with deep learnin
 g. Recently this approach has been pushed even further\, with Convolutiona
 l Neural Networks (CNNs)\, integrating strong inductive bias in their stru
 cture\, being replaced with Vision Transformers (ViT)\, or even fully-conn
 ected neural networks with alternative training regimes. In this talk\, Dr
 . Yanni Ioannou\, Assistant professor at the University of Calgary\, will 
 discuss the current limits of this approach\, and the need for more domain
 -agnostic learning.\n\nSpeaker(s): Yani Ioannou\, \n\nBldg: Virtual\, Calg
 ary\, Alberta\, Canada\, Virtual: https://events.vtools.ieee.org/m/308822
LOCATION:Bldg: Virtual\, Calgary\, Alberta\, Canada\, Virtual: https://even
 ts.vtools.ieee.org/m/308822
ORGANIZER:alireza.imani@ucalgary.ca
SEQUENCE:7
SUMMARY:Inductive Bias in Deep Learning: From Structure to Training
URL;VALUE=URI:https://events.vtools.ieee.org/m/308822
X-ALT-DESC:Description: <br /><p>Machine learning requires assumptions and 
 data\, with the assumptions based on domain knowledge. With the exponentia
 l increase in dataset set and computation we&rsquo\;ve been able to reduce
  the domain expertise and inductive bias we integrate in computer vision m
 odels\, from manually designed features such as SIFT\, to learning represe
 ntations with deep learning. Recently this approach has been pushed even f
 urther\, with Convolutional Neural Networks (CNNs)\, integrating strong in
 ductive bias in their structure\, being replaced with Vision Transformers 
 (ViT)\, or even fully-connected neural networks with alternative training 
 regimes. In this talk\, Dr. Yanni Ioannou\, Assistant professor at the Uni
 versity of Calgary\, will discuss the current limits of this approach\, an
 d the need for more domain-agnostic learning.</p>\n<div id="gtx-trans" sty
 le="position: absolute\; left: 981px\; top: 77.5625px\;">\n<div class="gtx
 -trans-icon">&nbsp\;</div>\n</div>
END:VEVENT
END:VCALENDAR