1,721,056 research outputs found
Proactive Persistent Agents - Using Situational Intelligence to Create Support Characters in Character-Centric Computer Games
No full textThroughout the 1990\u27s computer game development was dominated by the improvement of game graphics. However, graphical excellence is now the norm rather than the exception, and so other technologies are coming to the fore as developers strive to make their games stand out in an ever more crowded market. Sophisticated artificial intelligence (AI) is one such technology which is now receiving a large share of game development effort as game designers attempt to create more believable computer controlled characters.Influenced by the notion of situational intelligence, this work presents the proactive persistent agent (PPA) architecture, an intelligent agent architecture that has been developed to drive the behaviours of non player support characters in character centric computer games. The key features of the architecture are that: PPA characters display believable behaviour across a diverse range of situations, through the use of role passing and fuzzy cognitive maps; PPA based characters are capable of sophisticated social behaviours based on psychological models of personality, mood and interpersonal relationships; The architecture performs in real time on machines of modest specifications; It is relatively easy for non programmers to author characters behaviours using the architecture.The PPA architecture has been used in two demonstration applications (an adventure game and a system for animating sophisticated virtual humans) which will also be presented. These applications have been used to perform a rigorous evaluation of the architecture, which will also be described. Evaluation is notoriously difficult in systems used to simulate the behaviour of virtual characters, as the key notion of believability is troublingly subjective. However, in spite of the difficulties involved, an evaluation scheme has been developed and used. The evaluation shows that the PPA architecture is ideal for the contol of non player support characters in character centric games
Analysis of Spam
No full textSpam is a pervasive annoyance in the lives of the Internet user. It has exploded into all facets of communications from mobile phones to personal organisers, and it has become a topical subject of discussion due to recent media coverage. Spam has a tangible cost measured in lost productivity, bandwidth usage, administration, and invasion of privacy. As a result, an anti-spam industry has evolved in order to counter the spam attack with a focus on two spam-filtering categories: collaborative techniques and content-analysis techniques.
This research involved analysing a wide variety of e-mail in order to produce a profile of spam and, more importantly, develop a profile of the spammer. A number of fundamental questions are answered, for example: are current definitions of spam adequate and if so, are they globally applicable? There was also an investigation to examine the possibility of a spammer successfully targeting e-mail to an individual or group of individuals.
In this study, honeypot accounts were created and positioned to receive spam. The outcome of this research is a definitive guide to spam, which will provide researchers and regular Internet users alike with knowledge that will aid them in the fight against spam and facilitate the improvement of spam filtering techniques
A Knowledge-Light Mechanism for Explanation in Case-Based Reasoning
No full textDecision support systems are currently achieving higher classification accuracies by using more
complex reasoning mechanisms. Examples of such mechanisms include support vector machines
and neural networks. However in spite of these increases in accuracy many decision support
systems are not accepted by users. In domains where there is a high cost associated with incorrect
classifications, such as medical domains, users are not always willing to accept a decision support
system\u27s classification without proper justification.
In every walk of life, from the home to the workplace, people use explanations all the time
to justify their opinions. Explanations can have many different forms depending on the context
in which they are used. Over the last few decades there has been a vast amount of research by
philosophers into the importance and the requirements of suitable explanations.
In spite of the importance of suitable explanations to justify an opinion, many decision support
systems fall sort of this requirement. Part of the reason for this is that in many types of decision
support systems it is often extremely difficult, if not impossible, to produce explanations. This
is particularly the case for black box systems such as support vector machines. However in other
systems such as rule-based systems, where the explanation can be in the form of a rule, the
explanations can often be complicated and result in confusion for users.
Alternatively Case-based Reasoning (CBR) Systems lend themselves naturally to producing
explanations. As the reasoning in CBR systems is performed on the most similar past case(s) to
a current problem, these similar cases can be used as an explanation for a classification. As these
similar cases are real past problems, they are generally easily understood by users.
It is our believe however, that in CBR systems, that there are more suitable cases to use as
an explanation than simply using the most similar cases. It is our belief that these more suitable
cases lie between the problem case and the perceived decision boundary. This results in the cases
forming an a fortiori argument. We describe a framework that we have developed for selecting
such cases.
We also believe that it is often not enough, regardless of the suitability, to just use a case as
an explanation. In our framework we included a mechanism for generating explanatory text that
can express why the case is suitable, or in some situations aspects of the case that may not be
suitable. This explanatory text can further assist users to decide if they agree with the opinion ofCBR system.
Based on the developed framework we implemented a decision support system for use in the
domain of Bronchiolitis, a viral infection that effects young children. This system was used and
evaluated in the Kern Medical Center, Bakersfield, California during their Bronchiolitis season
Intelligent Multicast Internet Radio
No full textAs the internet continues to mature static content will continue to give way to
dynamic, interactive multimedia content. This content will enrich all online media
while at the same time imposing heavy bandwidth usage on the underlying network
infrastructure. Multicast networking will undoubtedly increase in popularity to
relieve the congested networks.
Even as the scaling power of multicast networking approaches, users are burdened by
the amount of multimedia and have to spend countless hours searching for the
appropriate content. Recommender systems are a key way to personalize this content.
This thesis is aimed at developing a scalable streaming application over the current IP
multicast infrastructure and then using personalization for recommending and
customizing dynamic multimedia content.
The system uses a web based personalization system as the basis for user interaction
and data collection. The multimedia streams are delivered through either Multicast or
Unicast depending on each user\u27s capabilities. In short it is a completely scalable,
intelligent, audio streaming application
Example-Based Machine Translation: An Adaptation-Guided Retrieval Approach
No full textTranslation can be viewed as a problem-solving process where a source language text is transformed into its target language equivalent. A machine translation system, solving the problem from first-principles, requires more knowledge than has ever been successfully encoded in any system. An alternative approach is to reuse past translation experience encoded in a set of exemplars, or cases.
A case which is similar to the input problem will be retrieved and a solution produced by adapting
its target language component. This thesis advances the state of the art in example-based machine translation by proposing techniques for predicting the adaptation requirements of a retrieval episode. An Adaptation-Guided Retrieval policy increases the efficiency of the retriever, which will now search for adaptable cases, and relieves the knowledge-acquisition bottleneck of the adaptation component. A flexible case-storage scheme also allows all knowledge required for adaptation to be deduced from the case-base itself.
The first part of the thesis contrasts such a CBR-motivated approach with current EBMT systems
which are either data-intensive or knowledge-intensive. A new EBMT scheme is proposed in which the cases encode knowledge about their own reusability, determined by cross-linguistic mappings.
The information allows cases to be generalised carefully, to the degree that is necessitated by the
data. Linguistic and translational divergences | the obstacles to reusability | are investigated in
the domain of software-manual translation, and on this basis, a suitable case representation scheme is proposed.
The second and third parts of the thesis describe the on-line and o_-line processes of an EBMT
system in which the case-base is the only knowledge source. Cases are deduced from texts automatically, and at run-time, the matching and retrieval tasks exploit the adaptability information in the cases in order to maximise coverage without compromising on accuracy. The multi-tiered case representation scheme allows adaptation at the sub-sentential and word levels, when necessary. The general performance of the system is shown to degrade gracefully and to improve as the case-base size increases
Making Personalised Flight Recommendations using Implicit Feedback
No full textAs e-commerce has become more popular, the problem of information overload has come to the
fore. Recommender systems that reduce the information overload problem are becoming more
common. However, the problem with many recommender systems is that they are associated with a high cost of learning customer preferences (in terms of cognitive load). We describe the Personal Travel Assistant (PTA), a flight recommender application that uses case-based reasoning (CBR) to overcome these problems.
The PTA allows users to search multiple flights providers concurrently and recommends flights
based on their individual travel preferences. These preferences are implicitly learned from observations of user behaviour. When the user purchases a flight, the PTA uses the selection of a preferred flight to discover and refine the user\u27s overall travel preferences. These preferences are stored in a user-model as sets of cases representing their interactions, which are used to provide personalised recommendations.
The PTA makes recommendations taking into account the context in which the flights were
offered. It uses features from the request to determine this context, e.g. the duration of the trip.
We perform evaluations of contextual recommendations that support our view that user preferences change depending on the context of the session. We further improve recommendation accuracy by storing and personalising similarity measures in the user-model. The PTA alters the relative importance of features in the personal similarity measure based on implicit user feedback, e.g. increasing the importance of price at the cost of stop-over time in a multiple hop flight.
We also investigate cooperative components to extend our recommendation strategies. These
allow users to reuse the information learned from other users when they encounter new situations.
However, these techniques are not as successful as we had hoped. We discuss these components in relation to other work on collaborative recommendation and suggest that the standard approach is unsuited to the PTA\u27s context-based recommendation strategy.
The strength of CBR in the e-commerce domain stems from its reuse of the knowledge base
associated with a particular application. Since case data may be one aspect of a company\u27s
entire knowledge system, it is important to integrate case data easily within a company\u27s IT
infrastructure, providing in effect a case-based view on relevant portions of the company knowledge base. We describe CBML, an XML-based Case Mark-Up Language we have developed to facilitate such integration
Story Games and the OPIATE System: Using Case-Based Planning for Structuring Plots with an Expert Story Director Agent and Enacting them in a Socially Simulated Game World
No full textStorytelling in computer games has become a major selling point for new titles. With new
games integrating compelling storylines with simulated worlds, there is increasingly a
standard set of techniques used to tell a story in games, including cut scenes, story-based
missions, and the unlocking of new areas of exploration with successful goal completion.
To further integrate story telling with the game play of future games, this work draws on
storytelling methods in games, along with academic AI techniques and simulation
methods, to propose an innovative paradigm for storytelling in future games.
This thesis presents a new approach to creating game mechanics, utilising a number of
key concepts that result in an interaction scheme that engages a player with a story, while
allowing the player the freedom to interact with and alter that story as it happens. A story
director agent was developed that uses case-based planning of skeletal plot scripts,
modelled on Propp\u27s morphology, and the dynamic adaptation of these plans.
This agent was incorporated into a social simulation engine that a player interacts with
through controlling one of the characters therein. The story director and social simulation
are symbiotically linked, with a feedback mechanism that ensures plots are planned
consistently with the simulation.
The system was evaluated and analysed and has proved to represent a successful
storytelling paradigm, implementing a high level of player interactivity with plotting,
while providing an experience that takes the form of an organically whole storyline
A Multiple Autonomous Agent System for Negotiating Valued Information Exchange on the Web
No full textThe amount of detailed personal information being recorded electronically about individuals is
growing rapidly. The increasing commercial pervasiveness of the Internet is a major factor in
this growth. This information has a significant economic value yet it is only relatively recently
that data subjects have begun to expect and receive some reward for the surrender and/or use of
their information. It is considered likely that this trend will continue, producing a growing
uneasiness over the erosion of personal privacy. In time, legislation or acceptable practice may
oblige data users who wish to gather consumer information, to pay data subjects to provide that
information.
The mechanics and logistics of data subjects supplying their own information to data users
necessitates an automated system to carry out the transactions. Any such system has to be able
to cope with the widest variety possible of data formats and quantities. It should allow the data
sellers to specify at a very fine-grained level which information should be sold as what price. It
should allow data users to stipulate what type of `souls? they are seeking, how many and how
much they want to spend.
This dissertation describes a system called the Soul Seller System that attempts to provide such
facilities. The system is based on mobile agent technology called Aglets. Data subjects provide
information about themselves via a web page. This data is passed to a selling Aglet (a `soul
seller?) that then represents the data subject in transactions with the `soul buyers?, other Aglets
who each represent a data user seeking information. When initially created, both types of Aglets
dispatch themselves to a central Aglet server, the `marketplace? where soul buyers and sellers
can meet each other. A conversation is initiated where soul buyers tell soul sellers what
information they are seeking. The soul sellers respond with the price this information would
cost. If the price is considered suitable, money and information are exchanged. Both data
subjects and data users can track what transactions have been carried out in their name via a web
page.
While the Soul Seller System works well on a small scale, problems are experienced when
attempts are made to scale it up. This is due to the limit on the number of Aglets that can be run
on just one Aglet server. Two possible solutions include using a database to represent the soul
sellers or to permit multiple marketplaces thereby enlarging the capacity of the system. It is
concluded that basing the transactions exclusively on Aglets is useful when developing a
prototype but would not be appropriate for the full application version
Automatic Web Publishing
No full textResearch and Development departments of most important companies
are currently producing technical documents and papers as a result of their
own investigation projects. This type of documents normally develop concepts
and ideas that often lead to a contribution to knowledge (or possibly to simple
information) that may be interesting for other research projects in the
department or for further investigation. In that kind of environments, a system
for publishing and sharing information become a requirement for the success
of the department.
Although many approaches to solve this problem have already been
considered, such as shared system files and mailing lists, all of them offer to
the user limited services and require extra input information to know which
documents are of interest to each user. Besides, the growing demand for web
based interactive solutions and for personalization has made them become
obsolete.
This project will investigate the design of an automated system for the
publication of technical documents on the World Wide Web. Using a simple
interface, its distributed users will be able to publish the results of their
research, and, automatically, the system will make the new document
available for the rest of the users (once reviewed if necessary) by linking it
from appropriate document collections, by building searching indices to allow
to look for the document using keywords. The publication of a new document
will be announced to potentially interested readers by emailing a notification
to them; to decide who might be interested in a new document, the system will
keep information about interests and past interactions with the system for
each authorized user. Other features that the system will provide (if
necessary) are the possibility of feedbacking information to report authors and
publishing the same document in different formats
- …
