--
Data Mining and Knowledge Discovery community,
focusing on the latest research and applications.
Contributions are most welcome and should be emailed,
with a DESCRIPTIVE subject line (and a URL, when available) to (kdd@gte.com).
E-mail add/delete requests to (kdd-request@gte.com).
Nuggets frequency is approximately weekly.
Back issues of Nuggets, a catalog of S*i*ftware (data mining tools),
and a wealth of other information on Data Mining and Knowledge Discovery
is available at Knowledge Discovery Mine site, URL http://info.gte.com/~kdd.
-- Gregory Piatetsky-Shapiro (moderator)
********************* Official disclaimer ***********************************
* All opinions expressed herein are those of the writers (or the moderator) *
* and not necessarily of their respective employers (or GTE Laboratories) *
*****************************************************************************
~~~~~~~~~~~~ Quotable Quote ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
One customer concerned that he would void the warranty if he dissassembled
the hard drive by removing the screws, ... used a hacksaw instead
(from ComputerWorld July 1, 1996 alt.cw page)
Previous
1 Next Top
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
[I am including the following message from AAAI to all KDD-96 attendees.
-- GPS]
Content-Type: text/plain; charset='us-ascii'
Date: Wed, 17 Jul 1996 14:47:50 -0700
From: aziz@aaai.org
(Hasina N. Aziz)
Subject: kdd banquet
Dear KDD-96 Attendee:
The KDD-96 Conference Banquet is scheduled for Saturday, August 3,
at the Benson Hotel, 7:00pm-10:00pm. The invited speaker will be
Inderpal Bhandari, IBM TJ Watson Research Center. The title of his talk
is Advanced Scout: Data Mining and Knowledge Discovery in NBA Data.
Cost: $55.00/person.
Please notify us by July 22, 1996 if you plan to attend this event by
sending a message to kdd@aaai.org
or fax to 415-321-4457. Due to current
low registration for this event, it may be cancelled unless we receive
a significant response.
Sincerely,
Annette Eldredge
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Date: Tue, 16 Jul 96 20:17:00 +0200
From: Jean-Luc SIMONI (simoni@tabarly.saclay.cea.fr)
I'm doing research in the field of Data Mining from fulltext document.
I would like to know if can give me some informations on persons or projects
that work in the same field as me.
Thank you for your answer,
Your sincerly,
Jean-Luc
-----------------------------------
Jean-Luc SIMONI
tel : (33) (1) 69.08.98.11
email: simoni@tabarly.saclay.cea.fr
C.E.A
Saclay, FRance
[I am also interested in the question below --
if you have some info please reply to simoni@tabarly.saclay.cea.fr
and cc to
kdd@gte.com.
GPS]
Previous
2 Next Top
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Date: Thu, 18 Jul 1996 18:04:42 -0400
From: Gregory Piatetsky-Shapiro (gps0@gte.com)
Subject: June 10, 1996 Information Week Cover Story: Digging Too Deep?
http://server1.cmp.farm.barrnet.net/cgi-bin/cqcgi?CQ_SESSION_KEY=FVQGJKMVCZXX&CQ_QH=125255&CQ_CD=2&CQ_TXT=Y
[If this URL does not work, go to
http://techweb.cmp.com/techweb/iw/current/
and search for
Cover Story in Information Week archive for June 10, 1996. GPS]
June 10, 1996
Issue: 583
Section: Top Of The Week
Cover Story
-- Data Dilemma -- Data warehousing and mining are booming.
The government now wants to know what
companies are doing with all that data.
By John Foley
A slew of new data warehousing and mining products promises to
make it even easier for companies to sort through, analyze, and store
information so they can better understand their customers. But be
warned: The government is keeping a closer eye on the kinds of data
you stockpile and how you use it.
The issue is privacy. Corporations are compiling huge databases
of customer-specific information, some of which is sensitive. Acxiom
Corp., which manages and sells a massive store of demographic,
economic, and lifestyle information, says its greatest competition is
coming from corporations building their own customer data
warehouses. As more corporations build these strategic databases, the
pressure is on to manage them prudently.
Just last week, the Federal Trade Commission held a two-day
workshop to determine if the government needs to set guidelines on how
companies use the data they collect over the Internet (see
story, p. 16). Advocates are pressing proposals to
regulate the way marketers garner and use information, especially
about children. Legislation also is pending on Capitol Hill.
Although concern over how companies handle personal information
is not new, it's mounting with the hyper-growth of the World Wide Web
and what one analyst calls self-service databases on the
Web. Corporations are making more information available over the
Web-and skimming Web-usage data to get a better understanding of who
Web users are and what they are looking for.
'But the risk to a company taking the wrong stance in this area
is quite high,' says Daniel Jaye, chief technology officer with CMG
Direct Interactive in Wilmington, Mass., which is building a warehouse
of Web usage data for interactive, online marketing. One corporate
concern: privacy-conscious Web surfers boycotting Web sites and the
associated loss of advertising dollars. 'If I [as a consumer] don't
have confidence that I know how this information is being used, I'm
going to walk away from the Internet,' says Jack Krumholtz, a
Microsoft representative in Washington.
The FTC will hold off taking action for now, but the agency says
it may revisit the matter in six months. Richard Jones, VP of customer
satisfaction with credit-card issuer MasterCard International in
Purchase, N.Y., says that gives corporations a window of opportunity
to demonstrate they can be trusted with consumers' personal
information. 'The challenge is to raise consumers' comfort level by
making them aware why this information is being collected and how it
can benefit them,' Jones says.
MasterCard promotes system security, self-regulation, and
consumer education as the best ways to diffuse privacy concerns. It
has reason to get others on board: Last fall, the company launched
MasterCard OnLine, a 1.2-terabyte database of credit-card transaction
information that it makes available to 22,000 member companies for a
fee. The database contains account numbers, not individual names or
addresses.
But MasterCard's member banks have their own databases with the
names and addresses of credit-card holders and can match that
information with what MasterCard provides online. So while MasterCard
has developed its own privacy rules, the company stops short of
telling its members what to do with the data they collect. 'Our
members have a direct relationship with their customers,' Jones
says. 'MasterCard does not dictate what information is acceptable and
what is not.'
CMG Direct Interactive promises to keep its data-collection
practices beyond reproach. The company, for instance, keeps
individuals' Web usage and preference data separate from their
identification and pledges never to join the two. 'We've set a goal
that we could not violate somebody's privacy even if we wanted to,'
Jaye says.
All kinds of corporations are striving to understand what makes
a customer, or potential customer, tick. By using data warehousing and
mining to focus on consumer preferences and buying patterns, they can
save millions of dollars in marketing expenses and generate millions
more in new sales.
Meantime, there's lots of personal data for sale. Data warehouse
specialist Acxiom, for instance, has information on 95% of U.S.
households. Odds are Acxiom knows your name, telephone number,
approximate income, height, weight, education, buying habits, type of
car, and credit-card use. It also knows about your family
relationships, including whether you have children. Other companies
that specialize in collecting consumer data include Donnelly
Marketing, MetroMail, and Polk.
'The idea about maintaining more information about people is
that you bother them less,' says Charles Morgan, Acxiom's CEO, who
adds: 'The more someone knows about me, the less likely they are to
send me an offer about a sewing craft kit.'
Acxiom keeps a staggering four terabytes of data on its
computers and has a library of 500,000 magnetic tapes. It takes 20
mainframes and 100 servers to comb through it all. To avoid misusing
the information, Acxiom has devised a nine-point privacy policy that
covers everything from consumer rights to system
security and employee training. 'We're very careful about who we do
business with and who we allow to have access to information,' says
Morgan. However, the ultimate responsibility for how that data is used
rests with Acxiom's customers, which include AT&T, Allstate Insurance,
Citibank, Wal-Mart, and Wells Fargo.
Other companies are holding on to the data they collect. Coopers
& Lybrand, for instance, plans to start mining data to expand its
business and improve customer service. But chief technology officer
Rowan Snyder says that data will remain 'confidential, proprietary
information about our clients. It will be used only for internal
purposes.'
The trend among companies to better understand their
customers-and themselves-is driving the data warehousing/mining market
to new heights. NeoVista Solutions, Pilot Software, and Red Brick
Systems are just three of a dozen or so vendors set to announce new
products and technologies this week-feeding a market expected to
quadruple to $8.8 billion by 1998.
'I call it a shift to fact-based decision-making from
seat-of-the-pants decision-making,' says Christopher Erickson,
chairman of high-flying Red Brick Systems, a Los Gatos, Calif.,
company that will announce plans this week to integrate data mining
algorithms from DataMind Corp. with its own relational database
management system. The product, due in the fourth quarter, will be one
of the first to combine a data warehouse platform with data mining,
making it easier for companies to predict future outcomes based on
patterns discovered within historical data.
The warehouse-mining combination appeals to Mike Faracca,
director of IT with UCA&L in Buffalo, N.Y., a sales and support
outsourcing firm for technology companies. 'It's very close to the
top of my strategic planning agenda,' says Faracca. UCA&L has built a
20-Gbyte warehouse of data collected from client telephone calls,
voice-response systems, and fax traffic. Later this year, the company
plans to spin off data marts that put that informa- tion directly into
the hands of its high-tech customers, including IBM and
Microsoft. Longer-term plans call for 'closed loop' data marts that
pass sales leads on to UCA&L clients, then track the customer contact
through product sales and registration.
NeoVista Solutions in Cupertino, Calif., will announce this week
a set of data-mining engines and data-transformation software that
work with a variety of relational database management systems.
Pilot Software in Cambridge, Mass., will introduce Discovery
Server, data mining server software that works with Pilot's decision
support and online analytical processing products. The announcements
coincide with the Data Warehouse World show, June 11 to 13, in Santa
Clara, Calif.
A.J. Brown, VP of marketing with DataMind, says his company is
monitoring the FTC and other government activity to determine what
impact it might have on the sale of data warehousing and data mining
products. Government scrutiny could be a boost, he says. 'It
highlights the need to use better techniques and do a better job of
targeting customers,' Brown says.
But corporations had better be careful: A poorly managed data
warehouse could put them into the public relations doghouse.
With additional reporting by Mary E. Thyfault and Marianne Kolbasuk
McGee
Copyright * 1996 CMP Media Inc.
Previous
3 Next Top
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Date: Tue, 9 Jul 1996 16:19:47 -0700 (PDT)
From: Computational Finance (compfin@cse.ogi.edu)
Subject: Computational Finance Program at OGI
http://www.cse.ogi.edu/CompFin/
COMPUTATIONAL FINANCE at the Oregon Graduate Institute of
Science & Technology (OGI)
An Intensive 12-Month Concentration in the MS Programs of
Computer Science & Engineering (CSE)
Electrical Engineering (EE)
=====================================================================
Program Overview:
Today's technology has increased the level of technical proficiency
required in the financial markets. At one time, for example,
spreadsheet skills, pre-calculus, and a basic understanding of
financial instruments were sufficient to build practical asset and
derivative pricing tools. Today, however, leading-edge financial
institutions routinely use advanced analytical and numerical
techniques from engineering and computer science to create, price,
and manage risk for both new and established instruments.
Advances in computing technology now enable the widespread use of
sophisticated, computationally-intensive analysis techniques, the
real-time analysis of tick-by-tick financial market data, and the
real-time management of portfolios of hundreds or thousands of
securities. Furthermore, modern data analysis tools can consider
many variables simultaneously and can capture complicated and often
nonlinear inter-dependencies between variables. This has opened
up new modeling possibilities for portfolio management, asset
allocation, hedging, derivatives instruments, and decision making.
The strong demand within the financial industry for
technically-sophisticated graduates who are well versed in
state-of-the-art quantitative analysis and computing techniques is
addressed at OGI by an intensive 12 month Computational Finance
program. Unlike a standard two year MBA, the program is directed
at training scientists, engineers, and technically-oriented financial
professionals.
The program is offered as a concentration in both the Computer
Science and Engineering (CSE), and Electrical Engineering (EE)
departments. The program leads to a Master of Science degree in
Computer Science and Engineering (CSE track), or in Electrical
Engineering (EE track). Computational Finance courses are also
cross-listed in the Management of Science & Technology (MST)
program.
The Computational Finance concentrations feature a unique combination
of courses that provide a solid foundation in finance at a non-trivial,
quantitative level, plus training in the essential core knowledge
and skill sets of computer science or the information technology
subdiscipline of electrical engineering. These skills are essential
for advanced analysis of markets and for the development of
state-of-the-art investment analysis, trading, derivatives pricing,
and risk management systems.
The MS in CSE is ideal preparation for students interested in
securing positions in information systems in the financial industry,
while the MS in EE provides rigorous training for students interested
in pursuing careers as quantitative analysts at leading-edge
financial firms.
The curriculum is strongly project-oriented, using state-of-the-art
computing facilities and live/historical data from the world's
major financial markets provided by Dow Jones Telerate. Students
are trained in using high level numerical and analytical packages,
such as MATLAB, Mathematica, and SPlus, for analyzing and modeling
financial data.
OGI has established itself as a leading institution in research
and education in Computational Finance. Moreover, OGI has very
strong research programs in a number of areas that are highly
relevant for work in quantitative analysis and information systems
in the financial industry. These areas include signal processing,
neural networks and adaptive systems, machine learning, information
theory and coding, nonlinear dynamics, stochastic processes, software
engineering, object-oriented programming, database systems,
transaction processing, human-computer interaction, and spoken
language understanding..
-------------------------------------------------------------------
Admission Requirements
-------------------------------------------------------------------
Applications for entrance into the Computational Finance MS programs
for Fall Quarter 1996 (which begins on Monday, September 23) are
currently being considered as they are received. Enrollment in
the program is limited.
Admission requirements are the same as the general require-
ments of the institution. GRE scores are required for the
12-month concentration in Computational Finance, although
they can be waived under certain circumstances.
A candidate must hold a bachelor's degree in computer sci-
ence, engineering, mathematics, statistics, one of the bio-
logical or physical sciences, finance, econometrics, or one
of the quantitative social sciences. Candidates who hold advanced
degrees in these fields or who have experience in the financial
industry are also encouraged to apply.
----------------------------------------------------------------------
Contact Information
----------------------------------------------------------------------
For more information, contact
Program Information Admission Information
E-mail: CompFin@cse.ogi.edu
Betty Shannon, Academic
WWW: Coordinator
http://www.cse.ogi.edu/CompFin/
Computer Science and
Engineering Department
Oregon Graduate Institute
of Science and Technology
P.O.Box 91000
Portland, OR 97291-1000
E-mail:
academic@cse.ogi.edu
Phone: (503) 690-1255
Previous
4 Next Top
>~~~Siftware:~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Date: Thu, 11 Jul 96 21:49:32 +0200
From: erict@isoftfr.isoft.fr
(Eric)
Subject: Introducing ALICE
Dear Sir,
Please find here a commercial description of ALICE, our new Data Mining
package. Further technical information will be sent in the following weeks.
Regards,
Eric
____________________________________________________________________________
______________
ALICE
Data Mining & Predictive modeling
ALICE finds the vital information hidden in your data and helps you use it
for making better decisions.
As your business evolves, more and more data describing your operations and
decisions are accumulated. These data hold valuable decision-making
information. Freeing this information is the key to more accurate decisions
and increased performance.
ALICE is a Data Mining tool that gives you instant access to the valuable
information hidden in your data. It connects to your data base, explores the
selected file and automatically identifies significant patterns and
cause-and-effect relationships that explain your business processes. The
high-value information uncovered by ALICE is delivered in the form of
easy-to-grasp decision trees.
As simple as a spreadsheet, ALICE is designed to be used by mainstream
business users. It does not require any specific technical skill and makes
state of the art Data Mining techniques available to data owners of all
kinds. With ALICE, you can now build more reliable predictive models and
make better decisions more easily and faster than ever before.
You collect large amounts of data that describe your products' performance,
your customers' behavior, yours patients' health, your stocks' profitability
and more generally, all kinds of events and populations. ALICE can help you
explain the variety of their performance and behavior. It works in almost
any industry and provides simple answers to such questions as: 'Who is the
target audience for my product?', 'Which type of clients represent a credit
risk?', 'Which type of patients are subject to side effects when using our
drugs?', 'Which factors influence the quality of the output?', 'Which type
of customers are satisfied with the service we provide?'.
ALICE is extremely easy to use. Select your data source - spreadsheet, data
base, data warehouse, stat packs - choose the file you want explored, point
at the field you want analyzed and let ALICE work. ALICE mines your data and
uncovers the factors explaining the phenomenon under examination. Results
are displayed in the form of decision trees that make it easy to spot the
most interesting relationships as well as exceptional cases.
ALICE builds decision trees automatically. These trees are interactive and
allow you to combine your own knowledge with the knowledge extracted by the
built-in algorithms. You can manipulate and browse information freely, tag
any particular nodes of the tree and generate a all kinds of outputs - text
reports, graphs, charts, SQL queries, rules and predictive models - all with
one simple button click.
ALICE is powerful
=B7 ALICE Data Mining engine is based on state of the art Machine Learning
technologies which complexity is totally hidden.
=B7 ALICE builds decision trees automatically, using default or user defined
parameters : Tree depth, minimum number of records per node, statistical
significance threshold, representativity threshold. You may also build the
tree step-by-step using the interactive mode, and focus on the branches you
want to further explore.
=B7 ALICE handles all types of numeric and alphabetic data, and includes a
wide range of algorithms for providing maximum resistance to unknown values.
Amounts of data and number of fields supported only depend on available=
memory.
ALICE is easy to use
=B7 ALICE offers extensive decision trees edition, interpretation and
exploitation functions.
=B7 With one button click, you can force splits, merge nodes, develop or=
cut,
collapse or expand, display more or less statistical figures: ALICE enables
you to manipulate the information in a very intuitive manner.
=B7 Browse information freely as you point and click in the tree browser;=
set
color codes that will automatically highlight the most significant
relationships, stick labels and comments to the nodes and make tree easily
readable.
=B7 Point at one or more nodes and create instantly all a variety of=
outputs:
text reports, graphs, charts, SQL queries, rules and predictive models.
ALICE fits into your IT system
=B7 ALICE is an open Data Mining tool that easily fits into your existing
Information System.
=B7 ALICE directly accesses data held in the leading spreadsheets (Excel,
Lotus), DBMSs (Access, Foxpro, Paradox, Dbase), stat packs (SPSS, SAS) and
EISs (BUSINESSOBJECTS). It also connects to any ODBC compliant package.
=B7 ALICE is available in both English and French versions. It runs on PCs
under Windows NT/95/3.1. Minimum system requirements: Pentium-based PC, MS
Windows NT/95/3.1, 8 Mo RAM, 5 Mo HD.
____________________________________________________________________________
______________
MEET US AT THE KDD - AAAI 96 EXHIBITION
PORTLAND OREGON - USA
FROM AUGUST 2 TO AUGUST 8
****************************************************************************
Eric Tocatlian, erict@isoft.fr
ISoft SA
Sales & Marketing Manager Chemin de Moulon
Directeur commercial 91190 Gif sur Yvette
Voice +33 1 6941 2777 FRANCE
Fax +33 1 6941 2532 info@isoft.fr
****************************************************************************
Previous
5 Next Top
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Date: Tue, 16 Jul 96 20:06:33 +0200
From: erict@isoftfr.isoft.fr
(Eric)
Subject: ISoft/BO partnership
ISOFT ANNOUNCES DATA MINING PARTNERSHIP WITH BUSINESS OBJECTS
Leading Data Mining Tools Vendor partners with Business Objects to Provide
Customers with Link Between BUSINESSOBJECTS and AC=B2 & ALICE.
Paris, France -- May 20, 1996 -- ISoft, one of the world's leading provider
of Data Mining tools and solutions, today announced a partnership with
Business Objects. ISoft and Business Objects will provide mutual customers
with a link between the companies' two products that will allow users to
launch ISoft's products AC=B2 and ALICE from within BUSINESSOBJECTS. This
link is available immediately, and will be provided by Business Objects to
customers at no charge.
ISoft one-button integration within BUSINESS OBJECTS will now make it easier
than ever for users to combine the benefits of data mining with the powerful
query, reporting, and OLAP capabilities of BUSINESSOBJECTS. Data mining
extends customers' decision support capabilities by helping users to
discover hidden trends and relationships in their data and to make
predictions using that information.
ISoft's AC=B2 and ALICE are high profile data mining products for=
exploring
databases through interactive decision trees and creating queries, reports,
charts, and even rules for predictive models. Since its first release in
1990, AC=B2 has emerged in the European market as a highly successful data
mining tool on the Unix and PC platforms. Available in 1996, ALICE
introduces major breakthroughs in terms of user-friendliness for data mining
on the PC.
'We are pleased to be partnering with ISoft, and to offer our customers the
ability to connect to AC=B2 through BUSINESSOBJECTS,' said Bernard=
Liautaud,
president and CEO of Business Objects. 'Our partnership will allow
customers to take advantage of BUSINESSOBJECTS for integrated query,
reporting, and OLAP, and AC=B2 for further data mining functions such as
decision trees and predictive modeling.'
'AC=B2 and ALICE identify the pieces of information that are specific to the
user's problem. Once the data has been located, using BUSINESSOBJECTS will
dramatically enhance AC=B2 and ALICE users' ability to navigate within the
information,' said Herve Perdrix, president and CEO of ISoft. 'Combining
our two products makes for a much more efficient decision making process, as
we are now able to provide our customers with a comprehensive and integrated
decision support software environment.'
About ISoft
ISoft has been a major provider of data mining solutions for the past 6
years, providing leading edge tools and applications to the most demanding
data owners throughout Europe. The company's product, AC=B2 and ALICE=
enable
all a variety of decision makers to easily locate key information within
corporate databases and save time while improving decision accuracy. AC=B2'=
s
unique 'knowledge modeling' feature translates complex information into
easily understood graphical models. Launching ALICE in 1996, ISoft now
makes the power of data mining available to every desktop PC user.
ISoft's products are available from ISoft and a worldwide network of local
resellers and VARs, including Bull. Other partners and users include Mitre
Corporation (defense industry), Finaref-La Redoute (retail and mail order
business), and the French Bank La Caisse d'Epargne.
About Business Objects
Business Objects (NASDAQ:BOBJY) is the world's leading supplier of
integrated query, reporting, and OLAP tools. The company's flagship product,
BUSINESSOBJECTS, provides mainstream business users with access to
information stored in corporate databases, data warehouses, and packaged
applications. The company pioneered the market for business-intelligent
decision support tools in 1990 by introducing the first product to use a
'semantic layer' to map complex database schemas to a business
representation understandable by non-technical end users.
Business Objects generated over $60M in revenue in 1995, concluding its
fifth consecutive year of more than 100% revenue growth. BUSINESSOBJECTS has
been licensed in more than 50 countries to over 2,900 customer sites and
160,000 users worldwide. Business Objects partners with more than 350
leading third-party vendors. More information on Business Objects can be
found on the World Wide Web at http://www.businessobjects.com.
BUSINESSOBJECTS is a trademark of Business Objects S.A.
AC=B2 and ALICE d'ISoft are trademarks of ISoft S.A.
Contact:
Tracy Eiler or Miriam Standish
BUSINESS OBJECTS
408/973-9300
teiler@busobj.com
mstandis@busobj.com=09
Eric Tocatlian
ISoft
+33 1 6941 2777
erict@isoft.fr=09
Danielle Dawson
BLANC & OTUS
415/512-0500
ddawson@bando.com
Previous
6 Next Top
>~~~Meetings:~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Date: Tue, 16 Jul 1996 11:30:01 -0400
From: iimam@verdi.iisd.sra.com
(Ibrahim Imam)
Subject: AAAI-96 WS on IAA (Call For Participation)
Call For Participation
AAAI-96 International Workshop
on
Intelligent Adaptive Agents (IAA-96)
Sunday, August 4, 1996, Portland, Oregon
http://www.mli.gmu.edu/~iimam/iaa96/ws_list.html
In Conjunction With
the Thirteenth National Conference on Artificial Intelligence AAAI-96
Sponsored by the American Association for Ar tificial Intelligence (AAAI).
*************************
(FINAL PROGRAM)
SUNDAY AUGUST 4, 1996
8:30 am - 5:30 pm
*************************
8:30 am - 8:45 : Opening Remarks
Ibrahim F. Imam
*************************
8:45 - 9:30 : Invited Talk (45 mins.)
Session Chair: Yves Kodratoff
'Directing Improvisational Actors'
Barbara Hayes-Roth, Stanford University, USA
*************************
9:30 - 10:30 : Adaptation in Multi-Agent Environment I
Session Chair: Brian Gaines
9:30 - 9:55 : (25 mins.)
Adaptive Intelligent Vehicle Modules for Tactical Driving,
Rahul Sukthankar, Shumeet Baluja, John Hancock, Dean Pomerleau, and Charles
Thorpe, Carnegie Mellon University, USA
9:55 - 10:15 : (20 mins.)
Adaptation Using Cases in Cooperative Groups,
Thomas Haynes and Sandip Sen, The University of Tulsa, USA
*************************
10:15 - 10:30 : (15 mins.)
Session Chair: Sandip Sen
Discussion I
Subject: Multiagent Learning and Adaptation
*************************
10:30 - 10:50 : Break
*************************
10:50 - 11:50 : Intelligent Adaptation
Session Chair: Brad Whitehall
10:50 - 11:15 : (25 mins.)
Knowledge-Directed Adaptation in Multi-Level Agents,
John E. Laird, Douglas J. Pearson, University of Michigan, and Scott B.
Huffman, Price Waterhouse Technology Center, USA
11:15 - 11:35 : (20 mins.)
Adaptive Methodologies for Intelligent Agents,
Ibrahim F. Imam, SRA International, USA
11:35 - 11:50 : (15 mins.)
Dynamic Aspects of Statistical Classification,
G. Nakhaeizadeh, Daimler-Benz Forschung und Technik (DE), C.C. Taylor,
University of Leeds (UK), and G. Kunisch, University of Ulm (DE).
*************************
11:50 - 12:10 pm : (20 mins.)
Session Chair: John Laird
Discussion II
Subject: Learning Across Levels s
*************************
12:10 pm - 1:10 : Lunch
*************************
1:10 - 1:55 : Invited Talk (45 mins.)
Session Chair: George Tecuci
'Adaptive Interactions in Societies of Agents'
Brian Gaines, University of Calgary, Canada
*************************
1:55 - 3:30 : Information-Based Adaptive Agents
Session Chair: John Laird
1:55 - 2:20 : (25 mins.)
Autonomous and Adaptive Agents that Gather Information,
Daniela Rus, Robert Gray, and David Kotz, Dartmouth College, USA
2:20 - 2:45 : (25 mins.)
Intelligent Adaptive Information Agents,
Keith Decker, Karia Sycara, and Mike Williamson, Carnegie Mellon University, USA
2:45 - 3:05 : (20 mins.)
Sacrificing vs. Salvaging Coherence: An Issue For Adaptive Agents In
Information Navigation,
Kerstin Voigt, California State University at San Bernardino, USA
*************************
3:05 - 3:30 pm : (25 mins.)
Session Chair: Kerstin Voigt
Discussion III
Subject: Adaptation Maintains The Utility of Agents in Ch anging
Environments
*************************
3:30 - 3:50 : Break
*************************
3:50 - 4:50 : Planning and Modeling in Adaptive Agents
Session Chair: Rahul Sukthankar
3:50 - 4:15 : (25 mins.)
Learning Reliability Models of Other Agents in a Multiagent System,
Costas Tsatsoulis, University of Kansas, and Grace Yee, Lockheed Martin, USA
4:15 - 4:35 : (20 mins.)
Using Perception Information For Robot Planning and Execution,
Karen Z. Haigh and Manuela M. Veloso, Carnegie Mellon University, USA
4:35 - 4:50 : (15 mins.)
Cooperative Agents That Adapt for Seamless Messaging in Heterogeneous
Communication Networks,
Suhayya Abu-Hakima, Ramiro Liscano, and Roger Impey, National Research
Council of Canada, Canada
*************************
4:50 - 5:30 : (15 mins.)
Session Chair: Costas Tsatsoulis and Ibrahim Imam
Discussion IV
Evaluation of the workshop (5 to 10 min.s talks)
A list of speakers will be announced at the Workshop
*************************
For comments and suggestions, send an email to:
e-mail: iimam@verdi.iisd.sra.com
< /A>
Previous
7 Next Top
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
KDD Nuggets Index