Projects related to Chapter 5 - Logic
Project 5.1: Semantic Brokering
The aim is to implement an application similar to the apartment
renting example in the book chapter.
Basic project (it should occupy 2 students for 2-3 weeks).
The following tasks should be carried out:
- Select a topic in which a brokering activity is to be
carried out. Here brokering refers to the matching of offerings and
requirements.
- Build at RDFS ontology for the domain.
- Populate the ontology with offerings, expressed in RDF.
- Express the selection criteria using nonmonotonic rules.
- Run your rules with the RDF/RDFS information using an
engine such as DR-DEVICE
or DR-Prolog.
To do so, you will need to express the rules in the format
prescribed by these systems.
Advanced project (it should occupy 2-3 students for a term project).
The aim is to implement a brokering scenario in a multi-agent
environment. Apart from carrying out the steps described in the
basic project, project participants need, among others, to:
- Develop a basic understanding of brokering in multi-agent
environments by studying some relevant literature, e.g.
K. Sycara, S. Widoff, M. Klusch, and J. Lu. Larks, Dynamic
Matchmaking Among Heterogeneous Software Agents in Cyberspace,
Autonomous Agents and Multi-Agent Systems 5, 2 (2002): 173-203 (pdf).
G. Antoniou, T. Skylogiannis, A. Bikakis, and N. Bassiliades, A
Deductive Semantic Brokering System, In Proc. 9th International
Conference on Knowledge-Based Intelligent Information and
Engineering Systems. LNCS 3682, Springer 2005, 746-752.
- Choose and familiarize themselves with a multi-agent
system. We have made good experience with JADE.
- Decide on the precise messages to be exchanged between agents.
- Find out how to call remotely the inference engine to be used.
Project 5.2: Proof Layer
The aim of this project is to realize a proof layer, the vision of
which was briefly outlined in Chapter 1. A first remark is that
there is not *the* proof layer, but rather a proof layer for
each selected semantic web reasoning system (logic). Still, some
considerations are common to all such systems.
Two possible logical systems for which one could implement a proof
layer are:
- A simple monotonic rules language, such as Datalog (Horn logic
without function symbols). The reasoning tool
Mandarax could be used.
- A nonmonotonic rules language, as discussed in this chapter.
DR-DEVICE
or DR-Prolog could be used.
Basic project (it should occupy 2-3 students for 2 months).
The aim is to develop an interactive system that provides
explanation to the user. Important aspects to be addressed include:
- Decide how to extract relevant information from the
overall proof trace. Project participants could consult the
automated reasoning and logic programming literature for ideas.
- Define levels of granularity for representing proofs.
Should whole proofs be shown, or only meta-steps? These could then
be refined if the user questions a certain step.
- Ultimately, the "leaves" of a proof will be RDF facts,
or rules, or inference conditions used.
Advanced project (it should occupy 3-4 students for a term project).
The aim is to implement a proof layer in a multi-agent environment.That is,
requests and proof parts will be exchanged between agents. Additional
aspects to be considered include:
- Choose and familiarize themselves with a multi-agent
system. We have made good experience with JADE.
- Represent proofs in an XML language, ideally extending RuleML.
- Decide on the precise messages to be exchanged between agents.
Submit a problem.
|