<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<style type="text/css" style="display:none;"><!-- P {margin-top:0;margin-bottom:0;} --></style>
</head>
<body dir="ltr">
<div id="divtagdefaultwrapper" style="font-size: 12pt; color: rgb(0, 0, 0); font-family: Calibri, Helvetica, sans-serif, "EmojiFont", "Apple Color Emoji", "Segoe UI Emoji", NotoColorEmoji, "Segoe UI Symbol", "Android Emoji", EmojiSymbols;" dir="ltr">
We are happy to announce the first CVPR workshop on Automatic Understanding of Visual Advertisements:
<a href="https://na01.safelinks.protection.outlook.com/?url=http:%2F%2Fpeople.cs.pitt.edu%2F~kovashka%2Fads_workshop%2F&data=01%7C01%7Ckovashka%40cs.pitt.edu%7Cb6e5063691f04c7491f508d594fd37d1%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=hgvWuqooMvYbJyR5ZJlCAPTZ%2B75JSY3jOdrVrt100Fg%3D&reserved=0" target="_blank" rel="noopener noreferrer" id="LPlnk368898" previewremoved="true">
<span id="LPlnk368898">http://people.cs.pitt.edu/~kovashka/ads_workshop/</span></a>
<br>
<div style="color: rgb(0, 0, 0);">
<div dir="ltr">
<div id="x_divtagdefaultwrapper" dir="ltr" style="font-size: 12pt; color: rgb(0, 0, 0); font-family: Calibri, Helvetica, sans-serif, "EmojiFont", "Apple Color Emoji", "Segoe UI Emoji", NotoColorEmoji, "Segoe UI Symbol", "Android Emoji", EmojiSymbols;">
<div><br>
As part of the workshop, we are organizing a <b>challenge </b>where the goal is to predict which statements accurately describe the implicit/explicit message that an advertisement conveys, e.g. "I should buy this car because it is spacious", "I should drink
this coffee because it will keep me awake", "I should eat this ice-cream because it's made of natural ingredients", "I should not pollute because it's bad for the animals", or "I should care about safety because human lives are fragile." Inferring all of these
statements requires recognition of creatively portrayed objects, their relationships and connotations, and performing common-sense reasoning. We are hosting the challenge on EvalAI:
<a href="https://na01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fevalai.cloudcv.org%2Fweb%2Fchallenges%2Fchallenge-page%2F86%2Foverview&data=01%7C01%7Ckovashka%40cs.pitt.edu%7Cb6e5063691f04c7491f508d594fd37d1%7C9ef9f489e0a04eeb87cc3a526112fd0d%7C1&sdata=9P6hSWh%2FTi3ieS%2BG0Cr4AWR5haq9GXoocVGT%2BWZhTBQ%3D&reserved=0" target="_blank" rel="noopener noreferrer" id="LPlnk770082" previewremoved="true">
<span id="LPlnk770082">https://evalai.cloudcv.org/web/challenges/challenge-page/86/overview</span></a>
<br>
<br>
Understanding advertisements is an AI-complete task, and can only be solved with expertise from a
<b>large number of areas, </b>some of which do not at the surface seem to be related to ads. Thus, as another part of the workshop, we invite submission of
<b>2-page abstracts</b> (works in progress, previously published or unpublished work) in any of these areas:<br>
<br>
<ul style="margin-top:0; margin-bottom:0">
<li> Visual persuasion, politics and social media</li><li> Metaphors and rhetoric in images or language</li><li> Recognition of objects in non-photorealistic datasets and domain adaptation</li><li> Modeling physical forces</li><li> Abnormality and surprise detection</li><li> Humor</li><li> Knowledge representation for visual question answering</li><li> Motivations of humans</li><li> Visual attention and memorability</li><li> Sentiment</li><li> Analyzing infographics, charts and comics</li><li> Other areas related to understanding persuasive creative imagery </li></ul>
<br>
If you are interested in this or related topics, please participate in our challenge, submit your 2-page abstract, or reach out to the organizers about other ways to become involved!</div>
<br>
</div>
</div>
</div>
</div>
</body>
</html>