<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<font face="Times New Roman, Times, serif">V torek, 15. oktobra 2019
bo ob 13ih v Oranzni predavalnici (drugo nadstropje glavne stavbe)
290. Solomonov seminar. Oranzna predavalnica je v drugem
nadstropju glavne stavbe IJS na Jamovi 39. Posnetki preteklih
seminarjev so na <a class="moz-txt-link-freetext"
href="http://videolectures.net/solomon/">http://videolectures.net/solomon/</a><br>
<br>
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~<br>
Title: Intentional Action in a Computational Cognitive System<br>
</font>
<pre class="moz-quote-pre" wrap=""><font face="Times New Roman, Times, serif">Lecturer: Will Bridewell, US Naval Research Laboratory
Abstract:
The current success of artificial intelligence and machine learning has encouraged speculation about artificial agency that makes little contact with the capabilities of existing technology. Specifically, it is unclear what an agent would amount to without mental states or their simulacra, yet progress in the development of computationally realized beliefs, desires, intentions, and so on has been minimal. This talk covers research on a computational theory of intentional action. At the heart of the theory is the claim that part of what it means to intentionally carry out a task is to be oriented toward the world in a task-relevant way. That is, agents preferentially attend to aspects of their environment that are relevant to their current tasks, which enables them to act in a goal-directed manner while remaining responsive to unpredictable events. A corollary of the claim is that failure to appropriately attend may result in unintentional outcomes. After motivating this research agenda, I will describe a computational implementation of the approach in the ARCADIA framework and demonstrate its application in the domains of visual question answering and autonomous vehicles.
Recommended Readings:
* Bello, P., & Bridewell, W. (2017). There Is No Agency Without Attention. AI Magazine, 38(4), 27-34. <a class="moz-txt-link-freetext" href="https://doi.org/10.1609/aimag.v38i4.2742">https://doi.org/10.1609/aimag.v38i4.2742</a>
* Bridewell, W., & Bello, P. (2016). A Theory of Attention for Cognitive Systems. Advances in Cognitive Systems, 4.
<a class="moz-txt-link-freetext" href="http://www.cogsys.org/papers/ACS2016/Papers/Bridewell_Bello-ACS-2016.pdf">http://www.cogsys.org/papers/ACS2016/Papers/Bridewell_Bello-ACS-2016.pdf</a>
* Bridewell, W., Wasylyshyn, C., & Bello, P. (2018). Towards an Attention-Driven Model of Task Switching. Advances in Cognitive Systems, 6.
<a class="moz-txt-link-freetext" href="http://www.cogsys.org/papers/ACSvol6/papers/paper-6-7.pdf">http://www.cogsys.org/papers/ACSvol6/papers/paper-6-7.pdf</a>
Biographical Sketch:
Will Bridewell is a computer scientist at the U.S. Naval Research Laboratory. Formerly he was a research scientist at Stanford University and the Institute for the Study of Learning and Expertise. His current research investigates the relationship between attention and intentional action with a broader interest in computational theories of consciousness. He holds Ph.D. and M.S. degrees in computer science from University of Pittsburgh and B.S. degrees in psychology, mathematics, and computers science from Northern Kentucky University. He is the co-designer of the ARCADIA cognitive system and co-directs the ARCADIA research program.</font></pre>
<font face="Times New Roman, Times, serif"><br>
<br>
<br>
<br>
</font>
<pre class="moz-signature" cols="72">
</pre>
</body>
</html>