Main Page Files Productions Problem Spaces Operators Goal Hierarchy Groups Prod. By Type All
Go to the documentation for this file...
00001 ##! 00002 # @mainpage Symbolic Concept Acquisition 00003 # 00004 # @kernel 8 00005 # 00006 # <h2>Symbolic Concept Acquisition</h2></p> 00007 # 00008 # Symbolic Concept Acqusition (SCA) is a model of concept learning 00009 # implemented in Soar. SCA was developed by Craig Miller. The 00010 # version of SCA documented here has been updated for Soar8 by <a href="mailto:wrayre@acm.org">Bob Wray</a>. The minor modifications made to SCA for 00011 # compatibility with Soar8 are detailed in the documentation. This 00012 # documentation also includes a <a href="SCA_Overview.pdf">conceptual overview of SCA</a>, as well 00013 # as specific documentation of the Soar production code. 00014 # <p> 00015 # SCA has also been extended 00016 # to include knowledge to map novel or unknown values of features to known values. Novel 00017 # features are often 00018 # introduced in transfer tasks in psychological experiments. The 00019 # solution proposed in this version of SCA is to allow novel 00020 # values to be mapped to known values. This mapping process competes 00021 # with abstraction. The result is 00022 # that unknown values may be ignored (via abstraction) or mapped to 00023 # a related value. (<a href="SCA_Overview.pdf">more information</a>) 00024 #<p> 00025 # All the files needed for SCA are included in 00026 # <a href="../../sca.zip">sca.zip</a>. 00027 # To provide a "working model", 00028 # some additional files are included 00029 # that can be used to run a simple model and trace the action of 00030 # SCA. These files are located in the directory test-harness/. 00031 # The test-harness productions set up a training instance for 00032 # prediction (no feedback), then learning (feedback provided), then 00033 # halt the model. You can step through the execution of this 00034 # example to see how the abstraction process works, and the 00035 # differences between prediction and training in the model. 00036 # Several logfiles are included (links below) that also 00037 # illustrate SCA with this test-harness. 00038 #<p> 00039 # Additional information about SCA: 00040 # <ul> 00041 # <li>Miller, C. S. (1993). Modeling Concept Acquisition in the Context of a Unified Theory of Cognition. Ph.D. thesis, The University of Michigan. Also available as Technical Report CSE-TR-157-93. 00042 # <li>Miller, C. S. and Laird, J. E. (1996). Accounting for graded performance within a discrete search framework. Cognitive Science, 20, 499-537. (<a href ="http://facweb.cs.depaul.edu/cmiller/abs96.html">abstract</a>) 00043 #<li>Wray, R. E., and Chong, R. S. (2003). <a href="http://www.speakeasy.net/~wrayre/pubs/QuantitativeExplorationsWithSCA_WrayChong_ICCM03.pdf">Explorations of quantitative category learning with Symbolic Concept Acquisition</a>. Presented at 5th International Conference on Cognitive Modeling (ICCM). Bamberg, Germany. April. 00044 #<li><a href="SCA_Overview.pdf">A brief overview of SCA and its application to a transfer task</a> 00045 #</P> 00046 #</P> 00047 #<li><a href="../../sca.zip">sca.zip</a>: zip file of SCA productions and this documentation </a> 00048 #<li>An <a href="../logfile1-filtered/logfile1.html">annotated trace</a> of SCA on a single instance, showing prediction and training behavior, decisions, and production firings. 00049 #<li>Original <a href="../logs/logfile1.txt">watch level 5 trace</a> showing all changes to WM during the run, along with decisions and production firings (not annotated). 00050 #</P> 00051 #</ul> 00052 00053 ##! 00054 # @group soar8 00055 # 00056 # Changes to SCA in order to support Soar8: 00057 # <ul> 00058 # <li> Removed operator reconsider productions <br>(no operator terminations needed in Soar8) 00059 # <li> Removed non-operator indifferent preferences <br> (all non-operator preferences are default parallel in Soar). 00060 # </ul></P> 00061 # <br>Additional changes are documented in individual files and productions: 00062 00063 ##! 00064 # @group transfer-task 00065 # 00066 # map-attributes was added for a transfer task. This is compatible with 00067 # the remainder of SCA but not needed if you have no need for a transfer 00068 # task. See the <a href="SCA_Overview.pdf">SCA overview</a> for more 00069 # information about the transfer task model. 00070 00071 ##! 00072 # @group default-rules 00073 # 00074 # These productions are not part of SCA per se. They are included in almost 00075 # every Soar application, and thus are included here as well. 00076 00077 ##! 00078 # @group test-harness 00079 # 00080 # These files are not part of SCA. They are included in order to 00081 # provide a working prototype version of SCA. 00082 #</P> 00083 # Note that the test harness productions have deliberately <b>not</b> been 00084 # specified with a type. Therefore, their type will appear as "unknown" 00085 # in the documentation (and can be readily differentiated from all the SCA 00086 # productions, which are typed). 00087 00088 source abstract.soar 00089 source count-attributes.soar 00090 pushd elaborations 00091 source elaborations_source.soar 00092 popd 00093 source get-example.soar 00094 source map-attributes.soar 00095 source note-relevant-feature.soar 00096 source prediction.soar 00097 source prediction-task.soar 00098 source reverse.soar 00099 source state-no-change.soar 00100 pushd state-no-change 00101 source state-no-change_source.soar 00102 popd 00103 00104 # These files for the test harness (examples) only: 00105 pushd test-harness 00106 source halt.soar 00107 source test-initialization.soar 00108 source stand-alone-control.soar 00109 popd 00110 00111 00112 00113 00114