Download Document

ITCC 2005
April 4-6, 2005 Las Vegas, NV
Managing a Relational Database
with Intelligent Agents

Ira Rudowsky, Ph.D
Brooklyn College/ Dept. Computer and Information Science
Institute of Neural and Intelligent Systems
Mount Sinai School of Medicine
Study Participants:
Ira Rudowsky, Ph.D
Theodore Raphan, Ph.D
Olga Kulyba
Mikhail Kunin. Ph.D.
Dmitri Ogorodnikov
Introduction
Research in computational neuroscience
 Digitized streams of sampled analog and digital data
 Static Image
 Video Streams
Choice of Prototype Application for Interaction
With Database
 An existing application (VMF), which is used to display
digitized analog data for a small group of neuroscience
researchers doing experiments with human and animal
subjects.
 Data need to be stored in an organized manner so that
they can be shared and mined using VMF
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 2
VMF Analyzer Application
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 3
Relational Database Design
 Bold fields in each table represent the primary key,
which consists of one or more fields
 SUBJECT_ID in EXPERIMENT is a foreign key as it
relates to SUBJECT_ID in the SUBJECT table
 Links between foreign and primary keys enable search
engine to quickly reduce the search space for a query.
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 4
Flexible Querying
During the course of experimentation, new types of
apparatus are being developed, additional types of
subjects are being tested and different types of tests
are being devised and need to be easily included in
database .
It became necessary to use free-format table entries
as descriptors for experiments to maintain
manageable tables.
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 5
Flexible Querying
General comment field in the EXPERIMENT table
provides for:
 free-format searches
 scalability without adding overhead and complexity
 Avoids the cost and time of programming changes to
the database and application code
Can search using AND and OR logical connectors
to enhance the querying capability
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 6
Flexible Querying
Query and Result screens
Free Format Text String for Search
Results of Search
However, this has raised issues of maintaining
data quality when it is entered so that the records
are properly indexed for efficient searching.
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 7
Data Quality
 Data quality problems due to a wide range of errors ranging
from simple entry to issues of data quality, which could
impact data analysis and further research
 A number of software tools have been developed to address
problems of data quality.
 Data cleansing examining the database following data entry to
insure that errors in the data are minimized
Methods of data cleansing include eliminating duplicate records
as a result of merged databases
 Decomposing data into elemental parts and reassembling the
data and verifying them across records
 A more systematic technique represents the cleansing as a
directed acyclic graph of data transformations. The data
cleansing is performed as an iterative procedure that traverses
the graph
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 8
Data Quality
 An alternate approach for data cleansing using artificial
intelligence is to utilize an expert system engine, Java
Expert System Shell (JESS), together with a set of rules
for data cleansing
JESS is a rule engine and scripting environment written entirely
in Java. JESS enables the creation of software that has the
capacity to "reason" using knowledge supplied in the form of
declarative rules. The validation and verification stage, however,
requires human intervention.
All presently proposed solutions to the data
cleansing problem are generally “reactive” in the
sense that the data cleansing is performed after the
accumulation of large amounts of data.
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 9
Belief, Desire & Intention Agents
 Agent-oriented programming has the potential for efficiently
dealing with improving the data quality of free formatted
fields.
 Agent-based systems differ from classical rule-based expert
systems in that they behave autonomously and interact with
each other.
 Agent-based programming is based on representation of
agent properties in terms of belief, desire, and intention
(BDI).
 An agent is specified in terms of its capabilities (things the agent can
do)
 A set of initial beliefs
 A set of initial commitments (an agreement to perform a particular
action at a particular time)
 A set of commitment rules
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 10
BDI Agents
Capabilities are used by agents to decide
whether to adopt commitments
An agent will not adopt a commitment to
perform an action if the agent determines
that there is no possibility for performing
that action.
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 11
Belief, Desire, Intention
The set of commitment rules determines how the
agent acts.
 Each commitment rule contains a message condition, a
mental condition and an action.
 In order to determine whether such a rule fires, the
message condition is matched against the message the
agent has received and the mental condition is matched
against the beliefs of the agent. If the rule fires, the agent
becomes committed to performing the action.
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 12
Example of BDI Function




Agent A sends a commitment request in a message to Agent B.
Agent B accepts or rejects the request based on the details of the request, its behavioral rules, and current
mental model.
Agent B then sends a message to Agent A indicating acceptance or rejection of the request.
If Agent B accepts the request, it agrees to perform the requested action at the requested time if possible.
 Agent B may have committed to making an inquiry into a database on behalf of Agent A. Even if
Agent B has the ability to connect and query the database, it may not be possible at the specified
time due to inaccessibility of the database.
 Agent B monitors the execution of the query and reports to Agent A on the success or failure of the
commitment.
action request to B from A
B accepts or reject A’s request
If accepted, B attempts to perform A’s request
Environment
Agent B
Agent A
message of success or failure
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 13
JACKTM Intelligent Agents
JACKTM Agent Language from Agent
Oriented Software
Agent pursues its given goals (desires),
adopting the appropriate plans
(intentions) according to its current set
of data (beliefs)
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 14
TM
JACK
Intelligent Agents
Agent pursues its given goals (desires), adopting the
appropriate plans (intentions) according to its
current set of data (beliefs)
 Agent - used to define the behavior of an intelligent
software agent by specifying the following:
Events , both internal and external, that the agent is
prepared to handle
Events that the agent can post internally to be handled by
other plans
Events the agent can send externally to other agents
Plans that the agent can execute
BeliefSets that the agent can refer to
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 15
JACKTM Intelligent Agents
 Event – an agent is motivated to take action due to:
Internal stimuli – an event an agent sends to itself , usually
as a result of executing reasoning methods in plans
External stimuli – messages from other agents or percepts
that an agent receives from its own environment
Motivations – goals that the agent is committed to
achieving
Categories of events
 Normal – represent transient information that the agent
reacts to (location of ball in soccer game)
 Agent selects first applicable plan instance for the event
and executes only that plan
 BDI – goal directed i.e., agents commits to the desired
outcome, not the method chosen to achieve it
 Agent selects from a set of plans, reevaluates if fails
excluding failed plans from further consideration
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 16
JACKTM Intelligent Agents
 Plan Analogous to an agent’s functions – the instructions
the agent follows to try to achieve its goals and handle its
designated events
Each plan handles a single event, but multiple plans may
handle the same event
An agent can further discriminate between plans by
executing a plan’s relevant() method to determine
whether it is relevant for the instance of a given event
From the relevant plans, it can further decide which are
applicable by executing the plans context() method
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 17
JACKTM Intelligent Agents
 BeliefSet maintain an agent’s beliefs about the world
 Represents an agent’s beliefs using a tuple-based relational
model.
 tuples can be true, false or unknown thus the concepts of
“beliefs”
 ClosedWorld Relations – the tuples stored are believed to be
true, those not stored are assumed false
 Open World Relations – store both true and false tuples,
anything not stored is “unknown”
Events can be posted when changes are made to the
beliefset and thus initiate action within the agent based on a
change of beliefs
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 18
Database – Agent Interaction
(1 ) Program.class
When a record with
SubjectType mouse or
monkey is found, it
invokes the method
submitUpdateRequest() of
the Updater agent
(4) SendUpdateCommand.plan
invokes update() method of Update
(6) SendUpdateCommand.plan
@send method sends Update
event to the Monitor agent.
Selects relevant plan to execute
(UpdateMonkey or UpdateMouse)
(5) Update.event
the update method is executed
(7) UpdateMonkey.plan
Handles the Update event,
updates record in table
@reply messages
the Finished event
4 April 2005
(3) UpdateRequest.event
the request method is invoked
(2) Updater.agent
submitUpdateRequest
invokes the request()
method of the
UpdateRequest event
(8)Finished.event
the finished() method is
executed
(9)SendUpdateCommand.plan
the @wait_for method receives the
reply from UpdateMonkey and returns
control to Program
Managing a Relational Database with Intelligent Agents
Slide 19
Conclusion
A simple intelligent agent-database interaction
has been developed which can be extended to
consider more complex data entries.
 It was tested by deliberately introducing errors
in data entry and was found to be responsive in
correcting errors.
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 20
Thank you!
You’ve been a great audience!
Questions, comments???
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 21
Agents
public agent Updater extends Agent {
#handles external event UpdateRequest;
#sends event Update;
#uses plan SendUpdateCommand;
#posts event UpdateRequest ev;
public Updater(String name) { super(name);}
public void submitUpdateRequest(String monitor, String
expID, String stype){
postEventAndWait(ev.request(monitor, stype, expID));
}
}
public agent Monitor extends Agent {
#handles external event Update;
#sends event Finished;
#uses plan UpdateMonkey;
#uses plan UpdateMouse;
#posts event Update ev;
public Monitor(String name){ super(name);}
String stype, eid;
public void setVars (String s, String e) {stype = s; eid = e; }
}
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 22
Events
public event Update extends BDIMessageEvent
{
public String stype, eid;
#posted as
update (String s, String e) {stype = s; eid = e; }
}
public event UpdateRequest extends BDIGoalEvent {
public String monitor, stype, eid;
#posted as
request (String m, String s, String e){
monitor = m; stype = s; eid = e;
}
}
public event Finished extends BDIMessageEvent {
public String stype, eid;
#posted as
finished(String s, String e) {stype = s; eid = e;}
}
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 23
Plans
public plan SendUpdateCommand extends Plan {
#handles event UpdateRequest preqev;
#sends event Update ev;
#reasoning method
body()
{try
{
Update q = ev.update(preqev.stype, preqev.eid);
@send (preqev.monitor,q);
@wait_for(q.replied());
Finished response = (Finished) q.getReply();
System.out.println(agent.name()+" has been updated in
SendUpdateCommand "+response.eid);
}
catch (NullPointerException npe) {
System.out.println("preqev.eid "+preqev.eid);
}
}
}
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 24
Plans
public plan UpdateMonkey extends Plan {
#handles event Update handleUpdateEvent;
#sends event Finished fev;
#uses interface Monitor self;
}
static boolean relevant(Update evRef)
{ return ((evRef.stype != null) && (evRef.stype.length() > 0));
context(){ handleUpdateEvent.stype.equals("monkey"); }
#reasoning method
body()
{ self.setVars(handleUpdateEvent.stype, handleUpdateEvent.eid);
// OTHER JAVA AND SQL CODE
s.executeUpdate("UPDATE EXPERIMENT SET SUBJECTTYPE='animal‘
WHERE EXPID='" + handleUpdateEvent.eid + "'" );
}
@reply (handleUpdateEvent,fev.finished(self.stype, self.eid));
}
}
4 April 2005
Managing a Relational Database with Intelligent Agents
Slide 25