PROJECTS AT THE DISTRIBUTED ARTIFICIAL INTELLIGENCE LAB by oas1s

VIEWS: 6 PAGES: 2

									                     Bar-Ilan Research & Development Company Ltd.
                     “The Technology Transfer Company of Bar-Ilan University”
                     Bar-Ilan University Campus, Ramat-Gan 52900, Israel
                     Phone: 972-3-5318441        Fax: 972-3-5356088
                     http://www.biu.ac.il/birnd  e-mail:birnd@mail.biu.ac.il


PROJECTS AT THE DISTRIBUTED ARTIFICIAL INTELLIGENCE LAB

                                                             Prof. Sarit Kraus
                                                             Mathematics Department


1. Intelligent Multimedia Authoring tools for electronic Publishing
The goal of this project is the development of software tools that will radically
simplify and accelerate the authoring process for creation of electronic
presentation (e.g., electronic magazines) by automating the complex, time-
consuming task of multimedia content selection. We are developing a Content
Manager (CM). The CM receives the raw, media-rich data input to the system.
This input takes the form of separate or linked objects containing text, graphics,
audio and video. It evaluates the content of each object based on the author's
specifications and the interests of the consumers or users of the electronic
presentation. User feedback is collected by intelligent agents and aggregated to
form an integrated user profile which is submitted to the CM.


2. Heuristic Methods for Solving the Dynamic Parking Assignment
Problem

We consider the dynamic parking assignment problem in the center of large
cities. We have been examining two different approaches to the parking
problem. These two approaches emerge from two disciplines: Economics and
Physics. The first technique takes advantage of the well-known bidding
process. The second approach is to implement the physics-based multi-agent
system model. While the physics-based multi-agent system model relies on
theories in Physics the bidding process relies on theories in Economics and
Game theory.

3. Automated Agents for Dissemination Environments

Ongoing advances in communications including the proliferation of the Internet,
the development of wireless networks, and the impending availability of high
bandwidth links to the home, have fueled the development of a wide range new
information-centered applications. Many of these new applications involve data
dissemination, that is, the delivery of data from a set of producers to a larger set
of consumers. Examples of data-dissemination-based environments include
broadcasting environments, and push-based environment.

The most critical requirement of the broadcasting system is receiving feedback
from the users on their specific interest domains in order to broadcast
information that satisfies the users' needs. In order to enable the broadcasting
system to receive the feedback, we shall assume that there is a way to transmit
information from a user's PC to the broadcasting system by means of a low
communication capacity. In previous work, (for which a patent was applied), we
have esigned and developed a software agents' model and algorithms to
facilitate the receiving of feedback from the users by the broadcasting system
and introduced the concept of integrated users profiles which represents the
users' interests.

In the current research, we extend the model and the algorithms to provide
more sophisticated techniques to handle issues which arise from the dynamic
nature of the broadcasting environment, on the one hand, and its continuity, on
the other. We will study problems such as updating the integrated users profile
over time, preparation of a broadcasting program, maintaining the documents
collection and handling new documents.

4. Reaching Agreements More Effectively in Electronic Commerce:
Strategies and Learning Techniques

Negotiations and auctions are among the main mechanisms for reaching
agreements between entities. In order to negotiate or to bid successfully an
agent needs an appropriate strategy that utilizes information about its
opponents, the issue under discussion, and the environment. The objective of
the proposed research is the identification of strategies and learning techniques
for the development of automated agents that will be able to participate
successfully in auctions and negotiations.

Our main application domain is the buying and selling of items in electronic
commerce, a medium which is becoming more and more widespread. We
consider issues related to the establishment of electronic auction houses, such
as the choice of appropriate protocols. However, our main effort is in the
development of automated agents that will be able to participate in such
auctions on behalf of the buyers and sellers. We expect that the development of
such agents will make electronic commerce more profitable and more widely
used.

For further information please contact Dr. Frances Shalit:
email: shalitf1@mail.biu.ac.il Tel: 972-3-531 8441

								
To top