Home
Research Interests
Publications
Projects
PC&Reviews
Students
View Aliaksandr Birukou's profile on LinkedIn

Current Research Interests

My current research interests include science 2.0 (improving the way science works), patterns, scientific communities, culture transfer, and culture of communities.

e-Science, e-Research, Science 2.0

No matter what you call it, I am interested in applying lessons from Open Source Software Development and from Web 2.0 to Science, to make it better. In this line, we have derived patterns of scientific publications (Xu2010) to help to make the semantics of the scientific discourse in the paper explicit. We have also developed the prototype of Liquid Journals, a novel approach for dissemination of scientific knowledge in the Web era. In Liquid Journals publication is separated from the dissemination and the notion of scientific contribution is extended to include demo, experimental workflows, blogs, datasets. Preliminary results are available in (Baez2010). Another line of research, is discovering scientific communities based on information about scientific networks. Preliminary results are available in (Mussi2010). You can learn more about research in this line by visiting the website of the project.

Formal definition of culture

I am interested in providing an operative definition of culture, which can be used in modern information systems (especially on the Web) to help communities to use the systems easier. Preliminary results are available in (Birukou2010, Birukou2009, Birukou2009b).

Patterns

I am interested in any kind of patterns representing best practices (software patterns, design patterns, patterns for time management, patterns for organizing conferences, and so on). In particular, I am interested in patterns as the means for capturing implicit or tacit knowledge and would like to see how it can be transferred to people who need such knowledge (and are not aware of it). Preliminary results of our attempt to build a wiki and a composite service for selecting patterns are available in (Weiss2007, Birukou2009a). People really liked the related work from (Birukou2009a), therefore I have elaborated it in a separate survey on searching and selecting patterns (Birukou2010a), which I am planning to update periodically.

Recommendation systems

I am interested in the topic of recommendation systems, in particular, in recommending patterns (design patterns, software patterns, etc.) and scientific publications (and scientific knowledge objects in general). In this line we have developed the IC-Service (Birukou2007), a general-purpose domain-independent recommendation service, which implements the Implicit Culture ideas. It can be used as a Java library, web service, or as an EJB component. This library is going to be released as an open-source system. Basically, I am maintaining the project and take care of extensions provided by other people. Preliminary results of recommendation of scientific publications are available in (Birukou2006).

Past Research Interests

I have background in service-oriented computing, diversity-, culture- and community-aware recommendation systems, data mining, multi-agent systems, compliance management, queuing systems.

Implicit Culture Framework

The Implicit Culture (IC) Framework (Birukou2010, Birukou2009, Birukou2006c) is a framework for computer-human, or, more generally, computer-agent interactions, which supports, preserves, and encourages desired behavior of the members of communities. IC combines multi-agent systems (MAS) and data mining techniques to facilitate the transfer of implicit knowledge and to assist in decision making process. The Implicit Culture ideas have been introduced a couple of years ago, and in my thesis I elaborated them and provided a formal definition of culture, developed the IC Framework and tested it in several information systems described below.
Web page: http://www.dit.unitn.it/~implicit/

Implicit

Implicit (Birukou2010, Birukou2005a, Birukou2005) is a recommendation system for web search. It analyzes search user behavior to grasp the implicit knowledge of the community of people and use it to produce meaningful and useful recommendations.

IC-Patterns

IC-Patterns (Birukov2006a) is a system that facilitates the process of selecting software patterns within a community of developers. The recommendation module of the system has been developed using the IC-Service library (Birukou2007). In this project we collaborated with Michael Weiss from Carleton University, Canada.

Web service discovery

This system (Birukou2007a, Kokash2007) helps a developer of a service-oriented application to choose the service with suitable functional and non-functional properties, such as QoS, price, etc. To achieve this, developers automatically share the history of their searches and service invocations. The recommendation module of the system has been developed using the IC-Service library. In this project we collaborated with Vincenzo D'Andrea and Natallia Kokash from the University of Trento.

Knowledge transfer

During my internship at the Center for Computational Analysis of Social and Organizational Systems (CASOS) at Carnegie Mellon University I worked on the project with Internal Revenue Service (IRS), where ways of discouraging people from participation in illegal tax schemes have been studied. Several US cities have been modeled using US Census Bureau data and spreading of information about schemes within the cities has been simulated using Construct, a multi-agent network model of information diffusion in a co-evolving social and cultural context. The results of the simulations provide useful insights on which methods of information diffusion should be adopted and how it is correlated with socio-demographic profiles of the cities (Hirshman2008).

Queueing systems and multi-agent systems

Just out of passion I collaborated with my former colleagues in Belarusian State University to compute performance characteristics of a multi-agent system using queueing theory. My role was to help with formulating the problem, they did the computation. Results are available in (Lee2007, Birukou2006b). Previously I was dealing with retrial queues and queueing systems that allow group processing of requests. I was also developing algorithms for controlling such queues.
www.000webhost.com