<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=us-ascii">
<title>Re: [ACT-R-users] what does negative decay mean?</title>
</head>
<body>
<font face="Calibri, Verdana, Helvetica, Arial"><span style="font-size:14pt">Coty,<br>
<br>
We have used very fast decay rates for perceptual memories on our robot (using SS-RICS). This is because the robot is perceiving so much information, so quickly, and we need to get rid of most of it as quickly as possible. This also means that you need different
activation equations for different memory types. So we can set activation equations based on memory type. <br>
<br>
We have also looked at using an equation that will reverse the activation..so activation in the negative direction. We have been experimenting using this as an instantiation of negative reinforcement; so trying to get rid of a behavior based on punishment.
This is in cases where we are using retrieval as the condition to execute a behavior (not utility). This also helps to get around the problem of executing a goal over and over because the activation of a retrieval continues to rise.<br>
<br>
Troy.<br>
<br>
<br>
On 1/15/13 9:05 AM, "Cleotilde Gonzalez" <<a href="coty@cmu.edu">coty@cmu.edu</a>> wrote:<br>
<br>
</span></font>
<blockquote><font face="Calibri, Verdana, Helvetica, Arial"><font size="2"><span style="font-size:11pt">ACT-R friends,<br>
During explorations of the decay parameter in the ACT-R activation equation, we have wondered if any body has ever used a negative value or even zero, or even something different from the .5 default value. What would a negative value of decay mean? just
curiosity, I don't think we are going to go that far :-)<br>
<br>
Coty <br>
</span></font><span style="font-size:14pt"><br>
<hr align="CENTER" size="3" width="95%">
</span></font><span style="font-size:14pt"><font face="Consolas, Courier New, Courier">_______________________________________________<br>
ACT-R-users mailing list<br>
<a href="ACT-R-users@act-r.psy.cmu.edu">ACT-R-users@act-r.psy.cmu.edu</a><br>
<a href="http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users">http://act-r.psy.cmu.edu/mailman/listinfo/act-r-users</a><br>
</font></span></blockquote>
<span style="font-size:14pt"><font face="Consolas, Courier New, Courier"><br>
</font><font face="Calibri, Verdana, Helvetica, Arial">Troy D. Kelley<br>
RDRL-HRS-E<br>
Cognitive Robotics and Modeling Team Leader<br>
Human Research and Engineering Directorate<br>
U.S. Army Research Laboratory<br>
Aberdeen, MD 21005<br>
Phone 410-278-5869 or 410-278-6748<br>
Note my new email address: <a href="troy.d.kelley6.civ@mail.mil">troy.d.kelley6.civ@mail.mil</a><br>
</font></span>
</body>
</html>