Here are my projects for 2012/13 – Please contact me to discuss any one of them.
My research has one central theme – agency – and the research that I am interested in can be realised in two different ways:
- Application of theory to practice. The design, specification and realisation of computer systems that exhibit agency. These projects are technical and require an understanding of programming and mathematics (mostly statistics). Example outputs might be intelligent software applications, games or simulations that demonstrate intelligence or autonomy.
- Critical review of theory. These projects take problems and use the research literature to propose potential solutions. Solutions are often ‘frameworks’, but in most cases there will be some validation required, usually in the form of a simulation or small prototype. There is less emphasis upon software development than the projects in Category (1).
I do not engage in projects that require human participants, questionnaires and such like.
All of my project students are expected to publish their work in appropriate fora, and I actively support this.
Category 1 Projects
Semantic Search Engine
Conceptual Graphs (see http://conceptualgraphs.org/) have applications in all domains where knowledge is encoded into computer systems. One such application is for Business Rules (see also http://www.brcommunity.com/p-n014.php). This project will develop two contributions:
- A prototype database mechanism that stores rules and allows queries to be generated based upon graph projection, and
- A prototype semantic search engine that will mine and update the database with augmented knowledge as a result of user searches.
Research Focus: conceptual graphs, graph theory, search algorithms
Technical Focus: OO programming, business rules implementation, web/mobile architectures, databases (essential)
Gamification is an interesting concept that can be used to reinforce the learning of specified behavioural characteristics. This project will identify a set of leadership behaviours and translate them into a set of tasks for a social game. Behaviours will be monitored by the game, and in response to certain conditions, new tasks will be specified. NB. Don’t underestimate this project, it is quite challenging. It may be suitable for two students to work together, each producing two separate, complementary contributions.
Research Focus: game theory, social networks, energy efficiency
Technical Focus: OO programming (Android/iOS), mobile HCI/UI design, HTML5, databases
The ‘wisdom of the crowd’ (http://en.wikipedia.org/wiki/Wisdom_of_the_crowd) is an interesting topic that lends itself to the modelling, analysis and evaluation of large collections of inter-cooperating agents in a Multi-Agent System (MAS). This project will utilise an existing problem domain to build a simulator that can be used to investigate the collective impact of experimental variables for singular agents or communities of agents, that are sub-sets of a much larger population.
Research Focus: social network analysis, simulation, MAS and agency, swarm, flocking behaviour
Technical Focus: OO programming (possibly), NetLogo (or similar), swarm toolkits, simulation building
Category 2 Projects
Collective Intelligence (CI) is a general term for the description of intelligence that emerges from the many interactions and collaborations between agents. This project will conduct a thorough and critical review of the CI literature and propose a solution for a known problem in the Cloud Computing domain. The proposal will make use of a simulation tool to produce a set of results that can be evaluated in the context of the literature review. If conducted properly, this project is likely to form the basis of a journal article or book chapter, to be published externally.
Research Focus: intelligence, agency, multi-agent systems, collective intelligence, cloud architectures
Technical Focus: simulation building
Peer-to-Peer (P2P) systems, when large enough, can exhibit emergent collective intelligence. Using P2P, agency and CI theory, this project will identify crucial behavioural primitives that result in identifiable traits of leadership in systems. Leadership traits might manifest themselves as networks that can proactively load balance, or optimised environments for enhanced resource discovery.
Research Focus: intelligence, agency, multi-agent systems, collective intelligence, P2P architectures
Technical Focus: simulation building