<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
On 2016-03-27 02:10, William Flynn Wallace wrote:<br>
<blockquote
cite="mid:CAO+xQEZJaGrk5-r7K2jc9ZHvPAVtZU=vBdx_iFSarPYGTJr3gw@mail.gmail.com"
type="cite">
<meta http-equiv="Content-Type" content="text/html;
charset=windows-1252">
<div dir="ltr">
<div class="gmail_default" style="font-family:'comic sans
ms',sans-serif;font-size:small;color:rgb(0,0,0)">
<blockquote
style="color:rgb(80,0,80);font-family:arial,sans-serif;font-size:12.8px;border-style:none
none none
solid;border-left-color:rgb(204,204,204);border-left-width:1pt;padding:0in
0in 0in 6pt;margin-left:4.8pt;margin-right:0in">
<p class="MsoNormal"><span
style="font-size:18pt;font-family:Calibri,sans-serif">I
think I am missing something here. You cannot mean a
literal translation of the visual or auditory field -
which should be identical. You want to experience the <i>interpretation
</i>of the field. The attitudes and so on, right?
Beautiful, ugly, boring, etc. Of course those would be
individual-specific, not gender-specific. <br>
</span></p>
</blockquote>
</div>
</div>
</blockquote>
<br>
Actually, that suggests an interesting augmented reality app.
Imagine running a computer vision system identifying things (we are
getting there for realtime captioning, <a class="moz-txt-link-freetext" href="https://vimeo.com/146492001">https://vimeo.com/146492001</a>,
it is not hard to imagine this working quite well in one or two
years). These things are run through a word meaning API (like
<a class="moz-txt-link-freetext" href="https://www.wordsapi.com">https://www.wordsapi.com</a> or <a class="moz-txt-link-abbreviated" href="http://www.wordnik.com">www.wordnik.com</a>) that gives a first
order interpretation. Now, the really interesting thing is that we
can do word association mining online or elsewhere to get
associations and sentiment - when you see something it could be
surrounded with associated attributes. If the associations are mined
from particular communities or experiences, then it might actually
give more sense of the interpretation of the field. A dollar bill
will have rather different halos when seen from a marxist, rapper,
poor, or libertarian context.<br>
<br>
(I got the idea from this webcomic strip,
<a class="moz-txt-link-freetext" href="http://www.sinfest.net/view.php?date=2011-10-09">http://www.sinfest.net/view.php?date=2011-10-09</a> )<br>
<br>
In practice this requires some not-quite-here yet tech to interpret
social roles and actions that I think current deep learning is still
bad at, but for concrete objects it shoud work. It actually sounds
like a project the MIT Media Lab or some similar institution might
do just for coolness. <br>
<br>
<br>
<pre class="moz-signature" cols="72">--
Anders Sandberg
Future of Humanity Institute
Oxford Martin School
Oxford University</pre>
</body>
</html>