http:^^www.cs.utexas.edu^users^estlin^pubs.html

This data set contains WWW-pages collected from computer science departments of various universities

MIME-Version: 1.0
Server: CERN/3.0
Date: Monday, 06-Jan-97 21:19:13 GMT
Content-Type: text/html
Content-Length: 8170
Last-Modified: Monday, 26-Aug-96 18:13:31 GMT

<title>Tara Estlin's Papers</title>

<h1>Tara Estlin's Papers</h1>

To view a paper, click on the open book image. <br> 
<br>


<ol>


<! ===========================================================================>

<a name="scope-aaai-96.ps.Z"</a>

<b><li>Multi-Strategy Learning of Search Control for Partial-Order
Planning<br></b>
Tara A. Estlin and Raymond J. Mooney<br>

<cite>Proceedings of the Thirteenth National Conference on Aritificial Intelligence</cite>,
pp. 843-848, Portland, OR, August, 1996. (AAAI-96)<p>

<blockquote>
Most research in planning and learning has involved linear,
state-based planners. This paper presents SCOPE, a system for learning
search-control rules that improve the performance of a partial-order
planner.  SCOPE integrates explanation-based and inductive learning
techniques to acquire control rules for a partial-order planner.
Learned rules are in the form of selection heuristics that help the
planner choose between competing plan refinements.  Specifically,
SCOPE learns domain-specific control rules for a version of the UCPOP
planning algorithm. The resulting system is shown to produce
significant speedup in two different planning domains.
</blockquote>

<a href="file://ftp.cs.utexas.edu/pub/mooney/papers/scope-aaai-96.ps.Z">
<img align=top src="/users/ml/paper.xbm"></a><p>

<! ===========================================================================>

<a name="scope-msl-96.ps.Z"</a>

<b><li> Integrating EBL and ILP to Acquire Control Rules for Planning<br></b>
Tara A. Estlin and Raymond J. Mooney<br>

<cite>Proceedings of the Third International Workshop on
Multi-Strategy Learning</cite>, pp. 271-279, Harpers Ferry, WV, May
1996. (MSL-96).<p>

<blockquote>
Most approaches to learning control information in planning systems
use explanation-based learning to generate control rules.
Unfortunately, EBL alone often produces overly complex rules that
actually decrease planning efficiency.  This paper presents a novel
learning approach for control knowledge acquisition that integrates
explanation-based learning with techniques from inductive logic
programming. EBL is used to constrain an inductive search for
selection heuristics that help a planner choose between competing plan
refinements. SCOPE is one of the few systems to address learning
control information in the newer partial-order planners.
Specifically, SCOPE learns domain-specific control rules for a version
of the UCPOP planning algorithm. The resulting system is shown to
produce significant speedup in two different planning domains.
</blockquote>

<a href="file://ftp.cs.utexas.edu/pub/mooney/papers/scope-msl-96.ps.Z">
<img align=top src="/users/ml/paper.xbm"></a><p>

<! ===========================================================================>

<a name="scope-proposal-96.ps.Z"</a>

<b><li> Integrating Explanation-Based and Inductive Learning Techniques
to Acquire Search-Control for Planning<br></b>
Tara A. Estlin<br>

Ph.D. proposal, Department of Computer Sciences, University of Texas
at Austin, 1996. <p>

<blockquote> 
Planning systems have become an important tool for automating a wide
variety of tasks. Control knowledge guides a planner to find solutions
quickly and is crucial for efficient planning in most domains.
Machine learning techniques enable a planning system to automatically
acquire domain-specific search-control knowledge for different
applications.  Past approaches to learning control information have
usually employed explanation-based learning (EBL) to generate control
rules.  Unfortunately, EBL alone often produces overly complex rules
that actually decrease rather than improve overall planning
efficiency.  This paper presents a novel learning approach for control
knowledge acquisition that integrates explanation-based learning with
techniques from inductive logic programming. In our learning system
SCOPE, EBL is used to constrain an inductive search for control
heuristics that help a planner choose between competing plan
refinements. SCOPE is one of the few systems to address learning
control information for newer, partial-order planners.  Specifically,
this proposal describes how SCOPE learns domain-specific control rules
for the UCPOP planning algorithm. The resulting system is shown to
produce significant speedup in two different planning domains, and to
be more effective than a pure EBL approach.<p>

Future research will be performed in three main areas.  First, SCOPE's
learning algorithm will be extended to include additional techniques
such as constructive induction and rule utility analysis.  Second,
SCOPE will be more thoroughly tested; several real-world planning
domains have been identified as possible testbeds, and more in-depth
comparisons will be drawn between SCOPE and other competing
approaches.  Third, SCOPE will be implemented in a different planning
system in order to test its portability to other planning algorithms.
This work should demonstrate that machine-learning techniques can be a
powerful tool in the quest for tractable real-world planning.<p>
</blockquote>

<a href="file://ftp.cs.utexas.edu/pub/mooney/papers/scope-proposal-96.ps.Z">
<img align=top src="/users/ml/paper.xbm"></a><p>

<! ===========================================================================>

<a name="dolphin-ewsp-95.ps.Z" </a>

<b> <li> Hybrid Learning of Search Control for Partial-Order Planning <br> </b>

Tara A. Estlin and Raymond J. Mooney <br>

<cite>New Directions in AI Planning</cite>, M. Ghallab and A. Milani, Eds,
IOS Press, 1996, pp. 129-140. <p>  

<blockquote>
This paper presents results on applying a version of the DOLPHIN
search-control learning system to speed up a partial-order planner.
DOLPHIN integrates explanation-based and inductive learning techniques
to acquire effective clause-selection rules for Prolog programs.  A
version of the UCPOP partial-order planning algorithm has been
implemented as a Prolog program and DOLPHIN used to automatically
learn domain-specific search control rules that help eliminate
backtracking. The resulting system is shown to produce significant
speedup in several planning domains.
</blockquote>

<a href="ftp/papers/dolphin-ewsp-95.ps.Z">
<img align=top src="/users/ml/paper.xbm"></a><p>

<! ===========================================================================>


<a name="ewsp95-ph.ps" </a>

<b> <li> Why Real-world Planning is Difficult: A Tale of Two Applications<br> </b>

Steve Chien, Randall Hill, Jr., XueMei Wang, Tara Estlin, Kristina Fayyad, and Helen Mortenson <br>

<cite>New Directions in AI Planning</cite>, M. Ghallab and A. Milani, Eds,
IOS Press, 1996, pp. 287-298. <p>  

<blockquote>
In this paper we describe a number of obstacles hampering the
application of planning technology to real-world problems, as
encountered in two real-world planning projects at JPL: MVP - a
planning system for automated generation of image processing
procedures; and LMCOA - an intelligent system for assistance in
antenna operations.  First, we describe how existing planning
representation must be enhanced to represent and reason about aspects
of plans besides goal achievement - resource usage, quality, execution
time, flexibility, and generality.  Second, planning systems must be
able to fit into a wide range of operational contexts - most planning
tasks cannot be completely automated, therefore at a minimum the plans
produced must be easily understandable and modifiable by the users.
In some cases the user must be intimately involved in the plan
construction process itself.  Third, planning systems must be able to
compare favorably in terms of software lifecycle costs to other means
of automation such as scripts or rule-based expert systems.  This
means that development of intelligent tools and environments to
facilitate knowledge acquisition, validation, and maintenance are of
prime importance.  We hope that our description and elucidation of
these issues will lead to increased work in these areas.
</blockquote>

<a href="ewsp95-ph.ps.Z">
<img align=top src="/users/ml/paper.xbm"></a><p>

<! ===========================================================================>