<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=Windows-1252">
<style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style>
</head>
<body dir="ltr">
<div style="font-family: Arial, Helvetica, sans-serif; font-size: 11pt; color: rgb(0, 0, 0);" class="elementToProof">
<span style="font-family: Roboto, sans-serif; font-size: 14px; color: rgb(51, 51, 51);"><span style="font-family:Arial,Helvetica,sans-serif; font-size:11pt"><span class="x_elementToProof x_ContentPasted0" style="box-sizing:border-box; font-family:Arial,Helvetica,sans-serif">We
introduce A View From Somewhere (AVFS)—a </span></span></span><span style="font-family: Arial, Helvetica, sans-serif; font-size: 14px; color: rgb(51, 51, 51);">dataset of 638,180 face similarity judgments over 4,921 faces. Each judgment corresponds to the</span><font face="Arial, Helvetica, sans-serif" style="font-size: inherit; font-style: inherit; font-variant-ligatures: inherit; font-variant-caps: inherit; font-weight: inherit; color: rgb(51, 51, 51);"><span style="font-size:14.6667px"><span class="x_elementToProof" style="font-family:Arial,Helvetica,sans-serif">
</span></span></font><font face="Roboto, sans-serif" style="font-size: inherit; font-style: inherit; font-variant-ligatures: inherit; font-variant-caps: inherit; font-weight: inherit; color: rgb(51, 51, 51);"><span class="x_elementToProof" style="font-size:14px; font-family:Arial,Helvetica,sans-serif">odd-one-out
(i.e., least similar) face in a triplet of faces and is accompanied by both the identifier and </span></font><span style="font-family: Arial, Helvetica, sans-serif; font-size: 11pt; color: rgb(51, 51, 51);"><span class="x_elementToProof x_ContentPasted0" style="box-sizing:border-box; font-family:Arial,Helvetica,sans-serif">demographic
attributes of the annotator who made the judgment.</span></span><br>
</div>
<div dir="ltr">
<p class="x_elementToProof" style="box-sizing: border-box; margin: 0px 0px 10px; font-family: Roboto, sans-serif; font-size: 14px; font-weight: 300; color: rgb(51, 51, 51); background-color: rgb(255, 255, 255);">
<span style="font-family:Arial,Helvetica,sans-serif; font-size:11pt"><strong class="x_elementToProof" style="box-sizing:border-box; font-weight:700"><br>
</strong></span></p>
<p class="x_elementToProof elementToProof" style="box-sizing: border-box; margin: 0px 0px 10px; font-family: Roboto, sans-serif; font-size: 14px; font-weight: 300; color: rgb(51, 51, 51); background-color: rgb(255, 255, 255);">
<span style="font-family:Arial,Helvetica,sans-serif; font-size:11pt"><strong class="x_elementToProof" style="box-sizing:border-box; font-weight:700">Abstract</strong></span></p>
<p class="x_elementToProof" style="box-sizing: border-box; margin: 0px 0px 10px; font-family: Roboto, sans-serif; font-size: 14px; color: rgb(51, 51, 51); background-color: rgb(255, 255, 255);">
<span style="font-family:Arial,Helvetica,sans-serif; font-size:11pt">Few datasets contain self-identified sensitive attributes, inferring attributes risks introducing additional biases, and collecting attributes can carry legal risks. Besides, categorical labels
can fail to reflect the continuous nature of human phenotypic diversity, making it difficult to compare the similarity between same-labeled faces. To address these issues,
</span><span style="font-family:Arial,Helvetica,sans-serif; font-size:11pt">we present A View From Somewhere (AVFS)—a dataset of 638,180 human judgments of face similarity</span><span style="font-family:Arial,Helvetica,sans-serif; font-size:11pt">. We demonstrate
the utility of AVFS for learning a continuous, low-dimensional embedding space aligned with human perception. Our embedding space, induced under a novel conditional framework, not only enables the accurate prediction of face similarity, but also provides a
human-interpretable decomposition of the dimensions used in the human decision-making process, and the importance distinct annotators place on each dimension. We additionally show the practicality of the dimensions for collecting continuous attributes, performing
classification, and comparing dataset attribute disparities.</span></p>
<p style="box-sizing: border-box; margin: 0px 0px 10px; font-family: Roboto, sans-serif; font-size: 14px; font-weight: 300; color: rgb(51, 51, 51); background-color: rgb(255, 255, 255);">
<span style="font-family:Arial,Helvetica,sans-serif; font-size:11pt"><strong style="box-sizing:border-box; font-weight:700">ICLR 2023 Paper</strong></span></p>
<p class="x_elementToProof" style="box-sizing: border-box; margin: 0px 0px 10px; font-family: Roboto, sans-serif; font-size: 14px; font-weight: 300; color: rgb(51, 51, 51); background-color: rgb(255, 255, 255);">
<a href="https://arxiv.org/abs/2303.17176" data-auth="NotApplicable" class="" data-loopstyle="link" style="box-sizing: border-box; text-decoration: none; color: rgb(66, 139, 202);"><span style="font-family:Arial,Helvetica,sans-serif; font-size:11pt">arxiv.org/abs/2303.17176</span></a></p>
<p class="x_elementToProof" style="box-sizing: border-box; margin: 0px 0px 10px; font-family: Roboto, sans-serif; font-size: 14px; font-weight: 300; color: rgb(51, 51, 51); background-color: rgb(255, 255, 255);">
<span style="font-family:Arial,Helvetica,sans-serif; font-size:11pt"><strong style="box-sizing:border-box; font-weight:700">Code + Data</strong></span></p>
<p class="x_elementToProof elementToProof" style="box-sizing: border-box; margin: 0px 0px 10px; font-family: Roboto, sans-serif; font-size: 14px; font-weight: 300; color: rgb(51, 51, 51); background-color: rgb(255, 255, 255);">
<a href="https://github.com/SonyResearch/a_view_from_somewhere" data-auth="NotApplicable" class="" data-loopstyle="link" style="box-sizing: border-box; text-decoration: none; color: rgb(66, 139, 202);"><span style="font-family:Arial,Helvetica,sans-serif; font-size:11pt">github.com/SonyResearch/a_view_from_somewhere</span></a></p>
<div class="x_elementToProof elementToProof" style="box-sizing: border-box; margin: 0px 0px 10px; font-family: Roboto, sans-serif; font-size: 14px; font-weight: 300; color: rgb(51, 51, 51); background-color: rgb(255, 255, 255);">
<div class="_Entity _EType_OWALinkPreview _EId_OWALinkPreview _EReadonly_1"></div>
</div>
</div>
</body>
</html>