|
|
|
|
|
"Ontology recapitulates philately"
-- anon
|
THIS ISSUE |
Volume 1,Number 14
November 18, 1996
Baltimore, MD
|
http://www.cs.umbc.edu/agents/agentnews/1996/14/
Modified Wednesday, 20-Nov-1996 21:00:50 EST
|
AGENT NEWS |
Software Agents Mailing List
agents@cs.umbc.edu
|
The Software Agents
Mailing List is devoted to the issues of software agents, personal
digital assistants, software robots, knowbots, intelligent interface
agents, etc. The list was begun in 1994 by Ray Johnson, then at the
Lockheed Palo Alto AI Center and now at Sun and moved to UMBC in the
Fall of 1996. The list is available in two forms -- the regular
AGENTS list and the AGENTS-DIGEST list. If you subscribe to the
regular agents list, you will get each message posted to the list
immediately and if you subscribe to the digest, you will get a
periodic digest message containing messages posted since the last
digest. To join the mailing list send a message to majordomo@cs.umbc.edu with
subscribe agents or subscribe agents-digest in the body
of the message. Messages sent to agents@cs.umbc.edu will be
immediately distributed to all of the subscriber to agents, added to
the hypermail
archive, and added to the current digest for later distribution to
the agents-digest subscribers.
11/18/96
|
Positions available
|
New agent-related
positions were posted by
Carnegie
Mellon University , Empirical
Media Corporation, Norwegian
University of Science and Technology , Franz,
inc.,
Siemens Corporate Research, Inc. , and Penn State
University.
|
AGENT TECHNOLOGY |
Identify and IDML
|
Identify is a system done by Emerge Consulting (Palo Alto) for
finding products and pages that use IDML to identify themselves. IDML is a
SGML-compliant markup language that gives Internet marketers and
publishers a standard way to identify themselves, their content, and
their products. Content creators insert IDML markup that describe the
contents of a page or objects on the page interspersed with the HTML
and content . Browsers will ignore the IDML but IDML-savvy indexing
agents can use it to better understand the content. Identify is the
site where IDML tags from the world over are indexed and
available. With Identify, users can find products and information
using the simple and powerful Identify engine. Users can query solid
information that the publishers provide themselves.
The current IDML language has just four tags (ID-PUBLISHER, ID-INFO,
ID-PRODUCT, ID-SYSTEM) that can be used to describe a page or object,
each of which has a handful of attributes. The big question, of
course, is how to manage the growth and acceptance of the underlying
ontologies.
10/19/96
|
JAM
|
The Columbia JAM
Project is building an infrastructure for launching Java-based
learning agents over network-based information systems that then spawn
learned "classifier agents". These classifier agents migrate to other
sites and are combined by "meta-learning". The resultant
"meta-classifier agents" can then migrate as well to harvest additional
knowledge from other agents. The particular application under study
involves fraud and intrusion detection in financial information
systems. This work is being done in collaboration with the Financial
Services Technology Consortium, a not-for-profit R&D organization
whose members include many of the nation's largest banks and
associated vendor community. 10/16/96 |
Microsoft Agent
ActiveX for interactive software agents
|
Microsoft Agent
is a set of software services that supports the presentation of
software agents (applications that operate on the user's behalf) as
interactive personalities within the Microsoft Windows interface. By
providing support for visual personalities, Microsoft Agent
facilitates a new form of user interaction known as a conversational
interface. A conversational interface attempts to leverage natural
aspects of human dialogue and social interaction (also known as a
social user interface), and make user interfaces more appealing and
approachable for a wider variety of users." "The conversational
interface approach facilitated by the Microsoft Agent services is not
a replacement for an application's conventional graphical user
interface design; it is an extension and enhancement of the existing
interactive modalities of the Windows interface. Microsoft Agent
services are not intended as an exclusive interface any more than a
mouse is a replacement for the keyboard. Character interaction can be
blended with the conventional interface components such as windows,
menus, and controls. Therefore, Microsoft Agent services can be used
to enhance the interface of an existing application or as the
exclusive interface." ( It looks like this is mainly support for
simple animation to accompany interactions driven by conventional
programs. -- ed) 10/28/96 |
AGENTS ON THE NET |
Off-line browsing agents
|
TechWeb has a page on off-line web browsing agents. Beyond Browsing--Offline
Web Agents by Joel T. Patz. "Tired of waiting for web pages to
download? Offline web agents are designed to log on, get the
information you need, and log off while you are performing other
tasks. They come in a variety of flavors, but the most popular are
designed to log onto a web site, roam through a user-definable set of
levels (that is, you control how many levels of links are followed)
and return the pages to your hard disk, where you can read them at
your leisure. ..." The page identifies seven systems -- Folio Web
Retriever 2.0, The PointCast Network 1.1, Smart Bookmarks 2.0, WebEx
1.01, WebWhacker 2.0, Freeloader 2.0, METZ Netriever 1.1 -- and
includes a head-to-head comparison and a feature chart.
10/15/96
|
Another agent named Bob
|
Customized searching goes beyond the net. Gerd Meissner, who
helped customize the German edition of Edupage, has developed a search
service called BOB,
The Human Search Engine, which combines searches of the Net with
searches beyond the Net, to help you when you're looking for such
things as: a special German saying? A bookstore in Bavaria? Or an old
pal's address in Paderborn? From the requests made by private users,
schools and non-profit organizations, the service chooses at least one
every week to research and answer via e-mail -- for free. U.S.
customers are welcome. The address is info@adline.de or
http://www.adline.de. [from Edupage, 22 October 1996] 10/23/96
|
Browser Buddy
|
Browser Buddy bu Softbots, Inc.is an Internet
software agent that does WWW page fetching according to your
specifications. It also helps you setup your URL collections so that
they are hierarchically organized and augmented with descriptive
notes. URL collections are saved in a "Choices" file for later use. To
specify what you want to fetch, you open a Choices file and "select"
URLs by double clicking on their labels in the URL area of the
display. 10/19/96
|
|
"The Angel turns the overwhelming Internet universe into a world
personalized specifically for you. Your personal Angel profiles you
and intelligently recommends places on the web that will interest
you. You view these recommendations on your own World Page, and you
can also store and access them in Angel Marks NetAngels'
state-of-the-art bookmarking database." 10/31/96
|
|
Bot Spot calls itself "the
spot for all the bots on the web." Its purpose is to inform users of
the latest bots and Internet agents available.10/31/96
|
AGENT
FUNDING |
DARPA HPKB program
|
The DARPA
High Performance Knowledge Bases (HPKB) program is aimed at
producing the technology needed to enable system developers to rapidly
construct very large knowledge-bases that provide comprehensive
coverage of topics of interest, are reusable by multiple applications
with diverse problem-solving strategies, and are maintainable in
rapidly changing environments. It is envisioned that the process for
constructing these large, comprehensive, reusable, and maintainable
knowledge bases would involve three major steps: building foundation
knowledge; acquiring domain knowledge; and efficient problem solving.
(BAA 96-43. CBD Reference: October 3, 1996. Close Date: December 2,
1996). |
DARPA Advanced Simulation program
|
DARPA's Advanced Simulation
Technology Thrust program is focused on modeling synthetic
military forces in simulated environments. Relevant AI/agent
technologies including multi-agent collaborative decision making,
collaboration/communication between humans and agents, action
selection and planning, behavioral "realism" (model effects of stress
and other factors on performance), adaptation and learning, etc.
Proposals are due approximately 30 days from the CBD announcement on
9/23/96. 10/15/96 |
AGENTS IN PRINT |
DAI in Medicine on the WWW
|
Artificial Intelligence in Medicine will have a special issue on Distributed
Artificial Intelligence in Medicine over the World Wide Web which
will include papers investigating aspects of the enabling potential of
the WWW to support the development of distributed AI applications in
Medicine, including distributed ai architectures in medicine over the
WWW and agent-based medical applications using the WWW. Papers should
be submitted by March 30th, 1997. 10/15/96 |
Automated Negotiation in State Oriented Domains
|
Zlotkin, G. and Rosenschein, J.S. (1996),
"Mechanisms for Automated Negotiation in State Oriented Domains",
Journal of AI Research, Volume 5, pages 163-238. Abstract: This paper
lays part of the groundwork for a domain theory of negotiation, that
is, a way of classifying interactions so that it is clear, given a
domain, which negotiation mechanisms and strategies are appropriate.
We define State Oriented Domains, a general category of interaction.
Necessary and sufficient conditions for cooperation are outlined. We
use the notion of worth in an altered definition of utility, thus
enabling agreements in a wider class of joint-goal reachable
situations. An approach is offered for conflict resolution, and it is
shown that even in a conflict situation, partial cooperative steps can
be taken by interacting agents (that is, agents in fundamental
conflict might still agree to cooperate up to a certain point). A
Unified Negotiation Protocol (UNP) is developed that can be used in
all types of encounters. It is shown that in certain borderline
cooperative situations, a partial cooperative agreement (i.e., one
that does not achieve all agents' goals) might be preferred by all
agents, even though there exists a rational agreement that would
achieve all their goals. Finally, we analyze cases where agents
have incomplete information on the goals and worth of other
agents. First we consider the case where agents' goals are private
information, and we analyze what goal declaration strategies the
agents might adopt to increase their utility. Then, we consider the
situation where the agents' goals (and therefore stand-alone costs)
are common knowledge, but the worth they attach to their goals is
private information. We introduce two mechanisms, one 'strict', the
other 'tolerant', and analyze their affects on the stability and
efficiency of negotiation outcomes. 10/29/96 |
[an error occurred while processing this directive]