I received my MSc in Computer Science from Wroclaw University of Science and Technology (WrUST), Poland, and an MSc in Software Engineering from the Blekinge Institute of Technology, Sweden, both in 2006. I was awarded my PhD in Computer Science in November 2009 from Wroclaw University of Science and Technology, and in the same year I was appointed a Senior Visiting Research Fellow at Bournemouth University (BU), where from 2010 I was a Lecturer in Informatics. I joined King’s in November 2011 as a Lecturer in Computer Science. In September 2015 I returned to Bournemouth University as a Principal Academic in Computing where I was also a Head of SMART Technology Research Group and a member of Data Science Initiative. In September 2017 I joined UTS as Associate Professor in Network Science.
Over the last few years I have collaborated with various commercial organizations and research groups. These research efforts resulted in publication of over 80 research papers in journals, books and conference proceedings. I have also been involved in several successful research proposals.
GRANTS AND PROJECTS
1. Data Science and Analytics Training and Engagement Services for Business, (08/2016 – 07/2017, budget £50k, Higher Education Innovation Fund, Co-I).
2. Grant for Grant in predictive analysis of complex networks – building a Network of Excellence and Programme of Work (01/03/2016 – 31/07/2016, budget £2,000, BU, PI); internal grant at Bournemouth University.
3. ENGINE: European research centre of Network intelliGence for INnovation Enhancement (06/2013 – 05/2017, budget 4.731 Mio. EUR, European Commision, lead at partner organisation). The goal of this project, coordinated by the Wroclaw University of Technology, is to create European research centre of Network intelliGence for INnovation Enhancement.
4. iCANS initiative – interdisciplinary Complex Adaptive Networks and Systems Theory and Applications (05/2012 – 07/2012, budget £2,000, KCL, PI) internal grant at King’s College London, part of EPSRC Bridging the Gaps Interdisciplinary Informatics grant, project leader.
The main goal of this initiative is to enhance cross-disciplinary research at KCL by recognising the links between different research groups and individual researchers in the area of complex networked systems as well as organising meetings and discussion panels to develop new ideas for joint research.
5. The Computational Intelligence Platform for Evolving and Robust Predictive Systems (INFER) project (07/2010–06/2014; budget 1.55 Mio. EUR, European Commission, BU Transfer of Knowledge Coordinator). Project was funded by the European Commission within the Marie Curie Industry and Academia Partnerships & Pathways (IAPP) programme.
Project partners are Evonik Industries from Germany, Research and Engineering Centre (REC) from Poland, and the Smart Technology Research Centre of Bournemouth University in the UK. The project focuses on pervasively adaptive software systems for the development of a modular computational INtelligence software platform For Evolving and Robust predictive systems applicable in various commercial settings and industries. The main innovation of the project is a novel type of environment in which the “fittest” predictive model for whatever purpose will emerge either autonomously or by user high-level goal-related assistance and feedback. I acted as the Transfer of Knowledge coordinator at Bournemouth University side and I was also involved in two research tasks: “Advanced software engineering” and “Complexity research”.
6. GRASP# – Groups, Relationships and Activities of Suspected Persons; Analiza otoczenia spolecznego oraz powi?za? sieciowych osób poszukiwanych i podejrzanych o popelnienie przest?pstwa (10/2009–10/2011, budget £283,000, Polish Ministry of Science and Higher Education, Co-I)
The research and developmental grant from the Polish Ministry of Science and Higher Education. I was a project co-investigator and a member of the project steering committee. The aim of the project was to investigate and analyse the social connections and characteristics of people accused and suspected of committing crime.
7. IT SOA – Service-Oriented Architectures; Nowe technologie informacyjne dla elektronicznej gospodarki i spolecze?stwa informacyjnego oparte na paradygmacie SOA (01/2009–12/2012, budget £8.480 Mio., Polish Ministry of Science and Higher Education, researcher)
The project was developed within the Regional Operational Programme Innovative Economy 2008-2013. I was a member of the project team that was responsible for development of the model and methods for decision support system in the service oriented knowledge utility system.
8. An individual research grant from the Polish Ministry of Science and Higher Education (09/2008–11/2009, budget £8,000, Polish Ministry of Science and Higher Education, PI)
The title of the grant was “A method for analysis of node position in the network of internet users”, number N516 264935. I was a principal investigator of this project.
9. SNAP – Social Network Analysis Platform (02/2008–01/2009, budget £2,000, WRUT, PI)
The grant obtained from the vice-chancellor of the Wroclaw University of Technology. I was the project manager of the “SNAP – Social Network Analysis Platform” project, which was developed by the members of the DaniE group and the purpose of which was to facilitate research on different large social networks.
10. Grant for grants (04/2008–09/2008, budget £15,000, Polish Ministry of Science and Higher Education, Co-I)
The grant that aimed at providing funds for the proposal preparation of “Advanced Methods in Collaborative Knowledge Acquisition and Processing” project proposal within the EU FP7 People programme.
11. Nature-inspired Smart Information Systems – Coordination Action project within EU FP6 (11/2005–01/2008, budget 1 Mio. EUR, European Commission, researcher)
I acted as a member of the Nature-inspired Data Technology (NiDT) focus group within this Network of Excellence. I attended the meetings of the network of excellence at Majorca (06/2006), Tenerife (11/2006), and Malta (11/2007).
Can supervise: YES
My main areas of research are complex networked systems, and analysis of their dynamics and evolution, as well as predictive, adaptive modelling of networked systems. I also recently started research in a new direction – the application of machine learning approaches to networked, dynamical systems. Perfect example of such systems is social network, a concept that we all know very well as each of us is a part of one global network. This network is created by people and the interactions between them. We constantly create connections both in the real world (at home, school, office) and in the rapidly growing online world (Facebook, YouTube, Twitter, Flickr). In my research I investigate those systems, their characteristics and how they change over time. Examples of very interesting questions worth investigating are e.g. what causes that when we work together we can achieve more than when we work individually (concepts known as collective intelligence and emerging behaviour) or what makes that some of the videos, pictures, stories spreads through social network so quickly (known as viral chains).
a) Intelligent analysis of large complex networks – the networks that are in the area of my interest are extracted from large datasets obtained from telecommunication companies (British Telecom plc – BT), e-mail servers (WUT, Enron), multimedia sharing systems (Flickr), etc. The first research on investigating and analysing social networks was conducted as part of the EU FP6 Coordination Action project on Nature-inspired Smart Information Systems (11/2005 – 01/2008) where I acted as a member of the Nature-inspired Data Technology focus group. I presented the research results at the NiSIS symposia at Majorca (06/2006), Tenerife (11/2006), and Malta (11/2007).
b) Dynamics and predictive modelling of complex networked systems – this is area that currently becomes the main field of my research efforts. The conducted research is concerned with discovering patterns in nodes’ behaviours and the interactions between them. The analysis of these patterns and their changes in time allows prediction of the future behaviour of nodes and their relations. One of the ways to model the network dynamics is the application of methods based on the molecular modelling concept and other physically-inspired methods. Another approach that I investigate is the application of machine learning methods to infer and predict the future structure and characteristics of network.
c) Network motifs method in social networks – Network motifs are small subgraphs that reflect local network topology and were shown to be useful for creating profiles that reveal several properties of the network. The outcomes of my research have revealed that motif analysis enables the effective investigation of both network structure and patterns of interactions between nodes within the network. In addition, the analysis of network motifs dynamics can be utilized in detecting and exploration of changes in complex network structures.
d) Multirelational social networks – these are the networks in which more than one type of relationship exists. Different types of relationships can emerge from various communication channels, i.e. based on each communication channel separate relation that can be also called a layer of a network is created. The relationships are extracted from the users activities and if in the system the knowledge about more than one kind of activity is gathered then more than one type of connection can be defined. Different layers can be also built upon various nature of the connections between users, e.g. co-workers, family members, friends. The systems that can be used in such analysis are the multimedia sharing systems such as Flickr or YouTube, which are typical examples of Web 2.0 systems. In my research I have investigated such systems as Flickr, Vimeo, ExtraDom, and recently Badoo.
e) Evaluation of a user position in a social network – during my PhD I conduct research on assessing a position of individuals in networked systems. The position is calculated based on the users activities and their interactions. The method and appropriate algorithms were developed and number of experiments on real-world networks was carried out. The research on evaluating user position was supported by the individual grant that I obtained from the Polish Ministry of Science and Higher Education (06/2008 – 12/2009).
f) Modelling of complex adaptive software systems – the main challenge of the current software systems is to build the systems that will be able to adapt to the changing external environment. The research conducted within the Advanced Software Engineering task within INFER project was focused on developing architectures for complex adaptive systems. I worked on that task when I was employed at BU.
I have taught different units at Undergraduate and Master levels and I have supervised several Undergraduate and Master Projects.
1. Bournemouth University, United Kingdom (2015-2017)
- Postgraduate Framework
2016/2017 – Advanced Data Management, 20 students, unit leader
- Undergraduate Framework
2015/2016 – Data Management, laboratories, 100 students
2015/2016 – Project Management and Teamwork, 180 students, unit leader
2. King’s College London, United Kingdom (2011-2015)
- Postgraduate Framework
2012/2013, 2013/2014 – Project Management, 105 students, module leader
- Undergraduate Framework
2011/2012, 2012/2013 – Data Structures, 180 students, module co-leader
3. Bournemouth University, United Kingdom (2010-2011)
- Postgraduate Framework (Master Course in Information Technology)
2010/2011 – Software Project Management, 13 students, unit leader
2010/2011 – Business System Design, 13 students, unit leader
- Undergraduate Framework
2010/2011 – Data Management, laboratories, 160 students
4. Wroclaw University of Technology, Poland (2006-2009)
- Master Framework (Master Course in Computer Science)
2008/2009 and 2009/2010 – Digital Image Processing, laboratories, lab leader, 45 students
2007/2008 – Data warehouses and data mining, laboratories, lab leader, 30 students
2007/2008 – Interactive web-based multimedia information systems design, laboratories, lab leader, 15 students
2007/2008 – Intelligent information systems, seminars, lab leader, 45 students
- Undergraduate Framework
2009/2010 – Databases, seminars, lab leader, 60 students
2007/2008 and 2008/2009 – Basics of Coding and cryptography, seminars, lab leader, 120 students
Abdulkareem, SA, Augustijn, E-W, Filatova, T, Musial, K & Mustafa, YT 2020, 'Risk perception and behavioral change during epidemics: Comparing models of individual and collective learning.', PLoS One, vol. 15, no. 1, pp. e0226483-e0226483.View/Download from: Publisher's site
Modern societies are exposed to a myriad of risks ranging from disease to natural hazards and technological disruptions. Exploring how the awareness of risk spreads and how it triggers a diffusion of coping strategies is prominent in the research agenda of various domains. It requires a deep understanding of how individuals perceive risks and communicate about the effectiveness of protective measures, highlighting learning and social interaction as the core mechanisms driving such processes. Methodological approaches that range from purely physics-based diffusion models to data-driven environmental methods rely on agent-based modeling to accommodate context-dependent learning and social interactions in a diffusion process. Mixing agent-based modeling with data-driven machine learning has become popularity. However, little attention has been paid to the role of intelligent learning in risk appraisal and protective decisions, whether used in an individual or a collective process. The differences between collective learning and individual learning have not been sufficiently explored in diffusion modeling in general and in agent-based models of socio-environmental systems in particular. To address this research gap, we explored the implications of intelligent learning on the gradient from individual to collective learning, using an agent-based model enhanced by machine learning. Our simulation experiments showed that individual intelligent judgement about risks and the selection of coping strategies by groups with majority votes were outperformed by leader-based groups and even individuals deciding alone. Social interactions appeared essential for both individual learning and group learning. The choice of how to represent social learning in an agent-based model could be driven by existing cultural and social norms prevalent in a modeled society.
Liu, X, Song, W, Musial, K, Zhao, X, Zuo, W & Yang, B 2020, 'Semi-supervised stochastic blockmodel for structure analysis of signed networks', Knowledge-Based Systems, vol. 195.View/Download from: Publisher's site
© 2020 Elsevier B.V. Finding hidden structural patterns is a critical problem for all types of networks, including signed networks. Among all of the methods for structural analysis of complex network, stochastic blockmodel (SBM) is an important research tool because it is flexible and can generate networks with many different types of structures. However, most existing SBM learning methods for signed networks are unsupervised, leading to poor performance in terms of finding hidden structural patterns, especially when handling noisy and sparse networks. Learning SBM in a semi-supervised way is a promising avenue for overcoming the above difficulty. In this type of model, a small number of labelled nodes and a large number of unlabelled nodes, coupled with their network structures, are simultaneously used to train SBM. We propose a novel semi-supervised signed stochastic blockmodel and its learning algorithm based on variational Bayesian inference, with the goal of discovering both assortative (the nodes connect more densely in same clusters than that in different clusters) and disassortative (the nodes link more sparsely in same clusters than that in different clusters) structures from signed networks. The proposed model is validated through a number of experiments wherein it compared with the state-of-the-art methods using both synthetic and real-world data. The carefully designed tests, allowing to account for different scenarios, show our method outperforms other approaches existing in this space. It is especially relevant in the case of noisy and sparse networks as they constitute the majority of the real-world networks.
Jin, D, Zhang, B, Song, Y, He, D, Feng, Z, Chen, S, Li, W & Musial, K 2020, 'ModMRF: A modularity-based Markov Random Field method for community detection', Neurocomputing, vol. 405, pp. 218-228.View/Download from: Publisher's site
© 2020 Elsevier B.V. Complex networks are widely used in the research of social and biological fields. Analyzing real community structure in networks is the key to the study of complex networks. Modularity optimization is one of the most popular techniques in community detection. However, due to its greedy characteristic, it leads to a large number of incorrect partitions and more communities than in reality. Existing methods use the modularity as a Hamiltonian at the finite temperature to solve the above problem. Nevertheless, modularity is not formalized as a statistical model in the method, which makes many statistical inference methods limited and cannot be used. Moreover, the method uses the sum-product version of belief propagation (BP) and its performance is not as good as the max-sum version, since it calculates per-variable marginal probabilities rather than the joint probability. To address these issues, we propose a novel Markov Random Field (MRF) method by formalizing modularity as an energy function based on the rich structures of MRF to represent properties and constraints of this problem, and use the max-sum BP to infer model parameters. In order to analyze our method and compare it with existing methods, we conducted experiments on both real-world and synthetic networks with ground-truth of communities, showing that the new method outperforms the state-of-the-art methods.
Verhoeven, D, Musial, K, Palmer, S, Taylor, S, Abidi, S, Zemaityte, V & Simpson, L 2020, 'Controlling for openness in the male-dominated collaborative networks of the global film industry.', PLoS One, vol. 15, no. 6, pp. e0234460-e0234460.View/Download from: Publisher's site
Studies of gender inequality in film industries have noted the persistence of male domination in creative roles (usually defined as director, producer, writer) and the slow pace of reform. Typical policy remedies are premised on aggregate counts of women as a proportion of overall industry participation. Network science offers an alternative way of identifying and proposing change mechanisms, as it puts emphasis on relationships instead of individuals. Preliminary work on applying network analysis to understand inequality in the film industry has been undertaken. However, in this study we offer a comprehensive approach that enables us to not only understand what inequality in the film industry looks like through the lens of network science but also how we can attempt to address this issue. We offer a data-driven simulation framework that investigates various what-if scenarios when it comes to network evolution. We then assess each of these scenarios with respect to its potential to address gender inequality in the film industry. As suggested by previous studies, inequality is exacerbated when industry networks are most closed. We review evidence from three different national film industries on network relationships in creative teams and identify a high proportion of men who only work with other men. In response to this observation, we test several mechanisms through which industry structures may generate higher levels of openness. Our results reveal that the most critical factor for improving network openness is not simply the statistical improvement of the number of women in a network, nor the removal of men who do not work with women. The most likely behavioural changes to a network will involve the production of connections between women and powerful men.
Naseem, U, Razzak, I, Musial, K & Imran, M 2020, 'Transformer based Deep Intelligent Contextual Embedding for Twitter sentiment analysis', Future Generation Computer Systems, vol. 113, pp. 58-69.View/Download from: Publisher's site
© 2020 Elsevier B.V. Along with the emergence of the Internet, the rapid development of handheld devices has democratized content creation due to the extensive use of social media and has resulted in an explosion of short informal texts. Although a sentiment analysis of these texts is valuable for many reasons, this task is often perceived as a challenge given that these texts are often short, informal, noisy, and rich in language ambiguities, such as polysemy. Moreover, most of the existing sentiment analysis methods are based on clean data. In this paper, we present DICET, a transformer-based method for sentiment analysis that encodes representation from a transformer and applies deep intelligent contextual embedding to enhance the quality of tweets by removing noise while taking word sentiments, polysemy, syntax, and semantic knowledge into account. We also use the bidirectional long- and short-term memory network to determine the sentiment of a tweet. To validate the performance of the proposed framework, we perform extensive experiments on three benchmark datasets, and results show that DICET considerably outperforms the state of the art in sentiment classification.
Wahid-Ul-Ashraf, A, Budka, M & Musial, K 2019, 'How to predict social relationships — Physics-inspired approach to link prediction', Physica A: Statistical Mechanics and its Applications, vol. 523, pp. 1110-1129.View/Download from: Publisher's site
© 2019 Elsevier B.V. Link prediction in social networks has a long history in complex network research area. The formation of links in networks has been approached by scientists from different backgrounds, ranging from physics to computer science. To predict the formation of new links, we consider measures which originate from network science and use them in the place of mass and distance within the formalism of Newton's Gravitational Law. The attraction force calculated in this way is treated as a proxy for the likelihood of link formation. In particular, we use three different measures of vertex centrality as mass, and 13 dissimilarity measures including shortest path and inverse Katz score in place of distance, leading to over 50 combinations that we evaluate empirically. Combining these through gravitational law allows us to couple popularity with similarity, two important characteristics for link prediction in social networks. Performance of our predictors is evaluated using Area Under the Precision–Recall Curve (AUC)for seven different real-world network datasets. The experiments demonstrate that this approach tends to outperform the setting in which vertex similarity measures like Katz are used on their own. Our approach also gives us the opportunity to combine network's global and local properties for predicting future or missing links. Our study shows that the use of the physical law which combines node importance with measures quantifying how distant the nodes are, is a promising research direction in social link prediction.
© 2018 Elsevier Inc. Change point detection in social networks is an important element in developing the understanding of dynamic systems. This complex and growing area of research has no clear guidelines on what methods to use or in which circumstances. This paper critically discusses several possible network metrics to be used for a change point detection problem and conducts an experimental, comparative analysis using the Enron and MIT networks. Bayesian change point detection analysis is conducted on different global graph metrics (Size, Density, Average Clustering Coefficient, Average Shortest Path) as well as metrics derived from the Hierarchical and Block models (Entropy, Edge Probability, No. of Communities, Hierarchy Level Membership). The results produced the posterior probability of a change point at weekly time intervals that were analysed against ground truth change points using precision and recall measures. Results suggest that computationally heavy generative models offer only slightly better results compared to some of the global graph metrics. The simplest metrics used in the experiments, i.e. nodes and links numbers, are the recommended choice for detecting overall structural changes.
Qin, M, Lei, K, Gabrys, B & Musial-Gabrys, K 2018, 'Adaptive community detection incorporating topology and content in social networks✰', Knowledge-Based Systems, vol. 161, pp. 342-356.View/Download from: Publisher's site
© 2018 In social network analysis, community detection is a basic step to understand the structure and function of networks. Some conventional community detection methods may have limited performance because they merely focus on the networks' topological structure. Besides topology, content information is another significant aspect of social networks. Although some state-of-the-art methods started to combine these two aspects of information for the sake of the improvement of community partitioning, they often assume that topology and content carry similar information. In fact, for some examples of social networks, the hidden characteristics of content may unexpectedly mismatch with topology. To better cope with such situations, we introduce a novel community detection method under the framework of non-negative matrix factorization (NMF). Our proposed method integrates topology as well as content of networks and has an adaptive parameter (with two variations) to effectively control the contribution of content with respect to the identified mismatch degree. Based on the disjoint community partition result, we also introduce an additional overlapping community discovery algorithm, so that our new method can meet the application requirements of both disjoint and overlapping community detection. The case study using real social networks shows that our new method can simultaneously obtain the community structures and their corresponding semantic description, which is helpful to understand the semantics of communities. Related performance evaluations on both artificial and real networks further indicate that our method outperforms some state-of-the-art methods while exhibiting more robust behavior when the mismatch between topology and content is observed.
Zhang, J, McBurney, P & Musial, K 2018, 'Convergence of trading strategies in continuous double auction markets with boundedly-rational networked traders', Review of Quantitative Finance and Accounting, vol. 50, no. 1, pp. 301-352.View/Download from: Publisher's site
© 2017, Springer Science+Business Media New York. This paper considers the convergence of trading strategies among artificial traders connected to one another in a social network and trading in a continuous double auction financial marketplace. Convergence is studied by means of an agent-based simulation model called the Social Network Artificial stoCk marKet model. Six different canonical network topologies (including no-network) are used to represent the possible connections between artificial traders. Traders learn from the trading experiences of their connected neighbours by means of reinforcement learning. The results show that the proportions of traders using particular trading strategies are eventually stable. Which strategies dominate in these stable states depends to some extent on the particular network topology of trader connections and the types of traders.
De Meo, P, Musial-Gabrys, K, Rosaci, D, Sarne, GML & Aroyo, L 2017, 'Using Centrality Measures to Predict Helpfulness-Based Reputation in Trust Networks', ACM TRANSACTIONS ON INTERNET TECHNOLOGY, vol. 17, no. 1.View/Download from: Publisher's site
Gao, F, Musial, K, Cooper, C & Tsoka, S 2015, 'Link Prediction Methods and Their Accuracy for Different Social Networks and Network Metrics', SCIENTIFIC PROGRAMMING, vol. 2015.View/Download from: Publisher's site
Budka, M, Juszczyszyn, K, Musial, K & Musial, A 2013, 'Molecular model of dynamic social network based on e-mail communication', Social Network Analysis and Mining, vol. 3, no. 3, pp. 543-563.View/Download from: Publisher's site
© 2013, The Author(s). In this work we consider an application of physically inspired sociodynamical model to the modelling of the evolution of email-based social network. Contrary to the standard approach of sociodynamics, which assumes expressing of system dynamics with heuristically defined simple rules, we postulate the inference of these rules from the real data and their application within a dynamic molecular model. We present how to embed the n-dimensional social space in Euclidean one. Then, inspired by the Lennard-Jones potential, we define a data-driven social potential function and apply the resultant force to a real e-mail communication network in a course of a molecular simulation, with network nodes taking on the role of interacting particles. We discuss all steps of the modelling process, from data preparation, through embedding and the molecular simulation itself, to transformation from the embedding space back to a graph structure. The conclusions, drawn from examining the resultant networks in stable, minimum-energy states, emphasize the role of the embedding process projecting the non–metric social graph into the Euclidean space, the significance of the unavoidable loss of information connected with this procedure and the resultant preservation of global rather than local properties of the initial network. We also argue applicability of our method to some classes of problems, while also signalling the areas which require further research in order to expand this applicability domain.
Musial, K, Budka, M & Blysz, W 2013, 'Understanding the Other Side - The Inside Story of the INFER Project', Smart Innovation, Systems and Technologies, vol. 18, pp. 1-9.View/Download from: Publisher's site
In the last few years, the collaboration between research institutions and industry has become a well established process. Transfer of Knowledge (ToK) is required to accelerate the development of both sides and to enable them to unlock their full potential. European Commission within the Marie Curie Industry and Academia Partnerships & Pathways (IAPP) programme supports the cooperation between these two sectors at the international scale by funding research projects that as one of the objectives aim at enhancing human mobility. IAPP projects offer people from different institutions the possibility to move sector and country in order to provide, absorb and implement new knowledge in a professional industrial-academic environment. In this paper, one of such projects is presented and both academia and industry perspectives in regard to opportunities and challenges in Transfer of Knowledge are described. Computational Intelligence Platform for Evolving and Robust Predictive Systems (INFER) is the IAPP project that serves as a case study for this paper. © Springer-Verlag Berlin Heidelberg 2013.
Musial, K, Budka, M & Juszczyszyn, K 2013, 'Creation and growth of online social network How do social networks evolve?', WORLD WIDE WEB-INTERNET AND WEB INFORMATION SYSTEMS, vol. 16, no. 4, pp. 421-447.View/Download from: Publisher's site
Brodka, P, Kazienko, P, Musial, K & Skibicki, K 2012, 'Analysis of Neighbourhoods in Multi-layered Dynamic Social Networks', INTERNATIONAL JOURNAL OF COMPUTATIONAL INTELLIGENCE SYSTEMS, vol. 5, no. 3, pp. 582-596.View/Download from: Publisher's site
Kazienko, P, Musial, K & Kajdanowicz, T 2011, 'Multidimensional Social Network in the Social Recommender System', IEEE TRANSACTIONS ON SYSTEMS MAN AND CYBERNETICS PART A-SYSTEMS AND HUMANS, vol. 41, no. 4, pp. 746-759.View/Download from: Publisher's site
Budka, M, Gabrys, B & Musial, K 2011, 'On Accuracy of PDF Divergence Estimators and Their Applicability to Representative Data Sampling', ENTROPY, vol. 13, no. 7, pp. 1229-1266.View/Download from: Publisher's site
Juszczyszyn, K, Kazienko, P & Musiał, K 2010, 'Personalized ontology-based recommender systems for multimedia objects', Studies in Computational Intelligence, vol. 289, pp. 275-292.View/Download from: Publisher's site
A framework for recommendation of multimedia objects based on processing of individual ontologies is proposed in the chapter. The recommendation process takes into account similarities calculated both between objects' and users' ontologies, which reflect the social and semantic features existing in the system. The ontologies, which are close to the current context, provide a list of suggestions presented to the user. Each user in the system possesses its own Personal Agent that performs all necessary online tasks. Personal Agents co-operate each other and enrich lists of possible recommendations. The system was developed for the use inthe Flickr multimedia sharing system. © 2010 Springer-Verlag Berlin Heidelberg.
Musiał, K, Juszczyszyn, K & Kazienko, P 2008, 'Ontology-based recommendation in multimedia sharing systems', Systems Science, vol. 34, no. 1, pp. 97-106.
In this paper, a new framework for recommendation of multimedia objects in the environment of the multimedia sharing system has been proposed. It uses two kinds of individual ontologies, one is created for multimedia objects and the second one for system users. The final recommendation process takes into account similarities calculated both between objects' and users' ontologies. These individual ontologies respect all the social and semantic features existing in the system. The entire recommender framework was developed for the use in Flickr, a typical photo sharing system.
Kazienko, P & Musial, K 2007, 'On utilising social networks to discover representatives of human communities', International Journal of Intelligent Information and Database Systems, vol. 1, no. 3-4, pp. 293-310.
: Virtual human communities that exist on the internet reflect social relationships between people. There is a great need to find important individuals, a set of people who would represent larger communities. These people would be able to perform specific tasks or could become a target group for marketing or advertising purposes. The new research on representative discovery for human communities is presented in this paper. Its main goal is to improve the process of target group selection by adding the social elements derived from the behaviours of people. The entire selection process considered in the paper is called human filtering. © 2007 Inderscience Enterprises Ltd.
Kazienko, P & Musial, K 2006, 'Recommendation framework for online social networks', ADVANCES IN WEB INTELLIGENCE AND DATA MINING, vol. 23, pp. 111-+.
Venkata, SK & Musial, K, 'Sub-query Fragmentation for Query Analysis and Data Caching in the Distributed Environment'.
When data stores and users are distributed geographically, it is essential to
organize distributed data cache points at ideal locations to minimize data
transfers. To answer this, we are developing an adaptive distributed data
caching framework that can identify suitable data chunks to cache and move
across a network of community cache locations.
Naseem, U & Musial-Gabrys, K 2019, 'Deep Intelligent Contextual Embedding for Twitter Sentiment Analysis', International Conference on Document Analysis and Recognition, IEEE, Sydney, Australia.View/Download from: Publisher's site
The sentiment analysis of the social media-based short text (e.g., Twitter messages) is very valuable for many good reasons, explored increasingly in different communities such as text analysis, social media analysis, and recommendation. However, it is challenging as tweet-like social media text is often short, informal and noisy, and involves language ambiguity such as polysemy. The existing sentiment analysis approaches are mainly for document and clean textual data. Accordingly, we propose a Deep Intelligent Contextual Embedding (DICE), which enhances the tweet quality by handling noises within contexts, and then integrates four embeddings to involve polysemy in context, semantics, syntax, and sentiment knowledge of words in a tweet. DICE is then fed to a Bi-directional Long Short Term Memory (BiLSTM) network with attention to determine the sentiment of a tweet. The experimental results show that our model outperforms several baselines of both classic classifiers and combinations of various word embedding models in the sentiment analysis of airline-related tweets.
Naseem, U & Musial, K 2019, 'DICE: Deep intelligent contextual embedding for twitter sentiment analysis', Proceedings of the International Conference on Document Analysis and Recognition, ICDAR, International Conference on Document Analysis and Recognition, IEEE, Sydney, Australia, pp. 953-958.View/Download from: Publisher's site
© 2019 IEEE. The sentiment analysis of the social media-based short text (e.g., Twitter messages) is very valuable for many good reasons, explored increasingly in different communities such as text analysis, social media analysis, and recommendation. However, it is challenging as tweet-like social media text is often short, informal and noisy, and involves language ambiguity such as polysemy. The existing sentiment analysis approaches are mainly for document and clean textual data. Accordingly, we propose a Deep Intelligent Contextual Embedding (DICE), which enhances the tweet quality by handling noises within contexts, and then integrates four embeddings to involve polysemy in context, semantics, syntax, and sentiment knowledge of words in a tweet. DICE is then fed to a Bi-directional Long Short Term Memory (BiLSTM) network with attention to determine the sentiment of a tweet. The experimental results show that our model outperforms several baselines of both classic classifiers and combinations of various word embedding models in the sentiment analysis of airline-related tweets.
Kitto, K, Sarathy, N, Gromov, A, Liu, M, Musial, K & Shum, SB 2020, 'Towards Skills-based Curriculum Analytics: Can we automate the recognition of prior learning?', LAK20: THE TENTH INTERNATIONAL CONFERENCE ON LEARNING ANALYTICS & KNOWLEDGE, 10th International Conference on Learning Analytics and Knowledge (LAK), ASSOC COMPUTING MACHINERY, ELECTR NETWORK, pp. 171-180.View/Download from: Publisher's site
Nguyen, TD, Maszczyk, T, Musial, K, Zöller, MA & Gabrys, B 2020, 'AVATAR - Machine Learning Pipeline Evaluation Using Surrogate Model', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 352-365.View/Download from: Publisher's site
© 2020, The Author(s). The evaluation of machine learning (ML) pipelines is essential during automatic ML pipeline composition and optimisation. The previous methods such as Bayesian-based and genetic-based optimisation, which are implemented in Auto-Weka, Auto-sklearn and TPOT, evaluate pipelines by executing them. Therefore, the pipeline composition and optimisation of these methods requires a tremendous amount of time that prevents them from exploring complex pipelines to find better predictive models. To further explore this research challenge, we have conducted experiments showing that many of the generated pipelines are invalid, and it is unnecessary to execute them to find out whether they are good pipelines. To address this issue, we propose a novel method to evaluate the validity of ML pipelines using a surrogate model (AVATAR). The AVATAR enables to accelerate automatic ML pipeline composition and optimisation by quickly ignoring invalid pipelines. Our experiments show that the AVATAR is more efficient in evaluating complex pipelines in comparison with the traditional evaluation approaches requiring their execution.
Wang, Y, Jin, D, Musial-Gabrys, K & Dang, J 2019, 'Community Detection in Social Networks Considering Topic Correlations', Proceedings of the AAAI Conference on Artificial Intelligence, AAAI Conference on Artificial Intelligence, AAAI, Hawaii, USA, pp. 321-328.View/Download from: Publisher's site
Network contents including node contents and edge contents can be utilized for community detection in social networks. Thus, the topic of each community can be extracted as its semantic information. A plethora of models integrating topic model and network topologies have been proposed. However, a key problem has not been resolved that is the semantic division of a community. Since the definition of community is based on topology, a community might involve several topics.
Wang, X, He, D, Jin, D, Musial, K, Liu, M & Dang, J 2019, 'Emotional contagion-based social sentiment mining in social networks by introducing network communities', International Conference on Information and Knowledge Management, Proceedings, ACM International Conference on Information and Knowledge Management, ACM, Beijing, China, pp. 1763-1772.View/Download from: Publisher's site
© 2019 Association for Computing Machinery. The rapid development of social media services has facilitated the communication of opinions through online news, blogs, microblogs, instant-messages, and so on. This article concentrates on the mining of readers' social sentiments evoked by social media materials. Existing methods are only applicable to a minority of social media like news portals with emotional voting information, while ignore the emotional contagion between writers and readers. However, incorporating such factors is challenging since the learned hidden variables would be very fuzzy (because of the short and noisy text in social networks). In this paper, we try to solve this problem by introducing a high-order network structure, i.e. communities. We first propose a new generative model called Community-Enhanced Social Sentiment Mining (CESSM), which 1) considers the emotional contagion between writers and readers to capture precise social sentiment, and 2) incorporates network communities to capture coherent topics. We then derive an inference algorithm based on Gibbs sampling. Empirical results show that, CESSM achieves significantly superior performance against the state-of-the-art techniques for text sentiment classification and interestingness in social sentiment mining.
Jin, D, Liu, Z, He, D, Gabrys, B & Musial, K 2018, 'Robust detection of communities with multi-semantics in large attributed networks', Knowledge Science, Engineering and Management 11th International Conference, KSEM 2018 Changchun, China, August 17–19, 2018 Proceedings (LNAI 11061), International Conference on Knowledge Science, Engineering and Management, Springer, Changchun, China, pp. 362-376.View/Download from: Publisher's site
© 2018, Springer Nature Switzerland AG. In this paper, we are interested in how to explore and utilize the relationship between network communities and semantic topics in order to find the strong explanatory communities robustly. First, the relationship between communities and topics displays different situations. For example, from the viewpoint of semantic mapping, their relationship can be one-to-one, one-to-many or many-to-one. But from the standpoint of underlying community structures, the relationship can be consistent, partially consistent or completely inconsistent. Second, it will be helpful to not only find communities more precise but also reveal the communities' semantics that shows the relationship between communities and topics. To better describe this relationship, we introduce the transition probability which is an important concept in Markov chain into a well-designed nonnegative matrix factorization framework. This new transition probability matrix with a suitable prior which plays the role of depicting the relationship between communities and topics can perform well in this task. To illustrate the effectiveness of the proposed new approach, we conduct some experiments on both synthetic and real networks. The results show that our new method is superior to baselines in accuracy. We finally conduct a case study analysis to validate the new method's strong interpretability to detected communities.
Wahid-Ul-Ashraf, A, Budka, M & Musial, K 2018, 'Netsim - The framework for complex network generator', Procedia Computer Science, Knowledge-Based and Intelligent Information & Engineering Systems, Elsevier, Belgrade, Serbia, pp. 547-556.View/Download from: Publisher's site
© 2018 The Author(s). Networks are everywhere and their many types, including social networks, the Internet, food webs etc., have been studied for the last few decadeS. However, in real-world networks, it's hard to find examples that can be easily comparable, i.e. have the same density or even number of nodes and edgeS. We propose a flexible and extensible Netsim framework to understand how properties in different types of networks change with varying number of edges and verticeS. Our approach enables to simulate three classical network models (random, small-world and scale-free) with easily adjustable model parameters and network size. To be able to compare different networks, for a single experimental setup we kept the number of edges and vertices fixed across the modelS. To understand how they change depending on the number of nodes and edges we ran over 30,000 simulations and analysed different network characteristics that cannot be derived analytically. Two of the main findings from the analysis are that the average shortest path does not change with the density of the scale-free network but changes for small-world and random networks; the apparent difference in mean betweenness centrality of the scale-free network compared with random and small-world networkS.
Abdulkareem, SA, Augustijn, EW, Musial, K, Mustafa, YT & Filatova, T 2018, 'The impact of social versus individual learning for agents' risk perception during epidemics', Proceedings - IEEE 14th International Conference on eScience, e-Science 2018, IEEE 14th International Conference on e-Science (e-Science), IEEE, Amsterdam, Netherlands, pp. 297-298.View/Download from: Publisher's site
© 2018 IEEE. Epidemics have always been a source of concern to people, both at the individual and government level. To fight outbreaks effectively, we need advanced tools that enable us to understand the factors that influence the spread of life-threatening diseases.
Musial-Gabrys, K, Wahid-Ul-Ashraf, A & Budka, M 2017, 'Newton's Gravitational Law for Link Prediction in Social Networks', Complex Networks & Their Applications VI Proceedings of Complex Networks 2017 (The Sixth International Conference on Complex Networks and Their Applications) (SCI 689), International Conference on Complex Networks and their Applications, Springer Verlag, Lyon, France, pp. 93-104.View/Download from: Publisher's site
Link prediction is an important research area in network science due to a wide range of real-world application. There are a number of link prediction methods. In the area of social networks, these methods are mostly inspired by social theory, such as having more mutual friends between two people in a social network platform entails higher probability of those two people becoming friends in the future. In this paper we take our inspiration from a different area, which is Newton's law of universal gravitation. Although this law deals with physical bodies, based on our intuition and empirical results we found that this could also work in networks, and especially in social networks. In order to apply this law, we had to endow nodes with the notion of mass and distance. While node importance could be considered as mass, the shortest path, path count, or inverse similarity (AdamicAdar, Katz score etc.) could be considered as distance. In our analysis, we have primarily used degree centrality to denote the mass of the nodes, while the lengths of shortest paths between them have been used as distances. In this study we compare the proposed link prediction approach to 7 other methods on 4 datasets from various domains. To this end, we use the ROC curves and the AUC measure to compare the methods. As the results show that our approach outperforms the other 7 methods on 2 out of the 4 datasets, we also discuss the potential reasons of the observed behaviour.
Butler, A, Xu, G & Musial-Gabrys, K 2018, 'Research Performance Reporting is Fallacious', International Conference on Behavioral, Economic, and Socio-Cultural Computing, IEEE, Taiwan, pp. 1-5.View/Download from: Publisher's site
Qin, M, Jin, D, He, D, Gabrys, B & Musial, K 2017, 'Adaptive community detection incorporating topology and content in social networks', Proceedings of the 2017 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, ASONAM 2017, pp. 675-682.View/Download from: Publisher's site
© 2017 Association for Computing Machinery. In social network analysis, community detection is a basic step to understand the structure, function and semantics of networks. Some conventional community detection methods may have limited performance because they merely focus on topological structure of networks. In addition to topology, content information is another significant aspect of social networks. Some state-of-the-art methods started to combine these two aspects of information, but they often assume that topology and content share the same characteristics. However, for some examples of social networks, content may mismatch with topological structure. In order to better cope with such situations, we introduce a novel community detection method under the framework of nonnegative matrix factorization (NMF). Our proposed method integrates topology and content of networks, and introduces a novel adaptive parameter for controlling the contribution of content with respect to the identified mismatch degree between the topological and content information. The case study using real social networks show that our new method can simultaneously obtain community partition and the corresponding semantic descriptions. Experiments on both artificial networks and real social networks further indicate that our method outperforms some state-of-the-art methods while exhibiting more robust behaviour when the mismatch topological and content information is observed.
Gao, F, Musial, K & Gabrys, B 2017, 'A community bridge boosting social network link prediction model', Proceedings of the 2017 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, ASONAM 2017, pp. 683-689.View/Download from: Publisher's site
© 2017 Association for Computing Machinery. Link prediction in social networks is a very challenging research problem. The majority of existing approaches are based on the assumption that a given network evolves following a single phenomenon, e.g."rich get richer" or"friend of my friend is my friend". However, dynamics of network dynamic changes over time and different parts of the network evolve in different manner. Because of that, we hypothesise that the prediction accuracy can be improved by providing different treatment to different nodes and links. Building on that assumption, we propose a Community Bridge Boosting Prediction Model (CBBPM) that treats certain bridge nodes differently depending on their structural position. For such bridge nodes their similarity score obtained using traditional link-based prediction methods is boosted. By doing so the importance of these nodes is increased and at the same time ensuring that the CBBPM can be used with any existing link prediction method. Our experimental results show that such bridge node similarity boosting mechanism can improve the accuracy of traditional link prediction methods.
Venkata, SK, Musial, K, Mahmoud, S & Keppens, J 2017, 'Multi-Agent System for Distributed Cache Maintenance', ADVANCES IN PRACTICAL APPLICATIONS OF CYBER-PHYSICAL MULTI-AGENT SYSTEMS: THE PAAMS COLLECTION, PAAMS 2017, 15th International Conference on Practical Applications of Agents and Multi-Agent Systems (PAAMS), SPRINGER INTERNATIONAL PUBLISHING AG, Porto, PORTUGAL, pp. 157-169.View/Download from: Publisher's site
Venkata, SK, Musial, K, Mahmoud, S & Keppens, J 2017, 'Demonstration: Multi-agent system for distributed cache maintenance', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 364-368.View/Download from: Publisher's site
© Springer International Publishing AG 2017. Innovations in science and technology is increasing the demand on huge data transfers and hence number of data caches. In this paper, we consider the community caching solution, CommCache, where many groups of users are working together on related projects distributed all over the world. We demonstrate the use of proactive caches for data placement problem with the help of multi-agent coordination.
Venkata, SK, Keppens, J & Musial, K 2015, 'Adaptive Caching Using Sub-query Fragmentation for Reduction in Data Transfers from Distributed Databases', ASTRONOMICAL DATA ANALYSIS SOFTWARE AND SYSTEMS XXV, 25th Annual Conference on Astronomical Data Analysis Software and Systems (ADASS XXV), ASTRONOMICAL SOC PACIFIC, ARC Ctr Excellence All Sky Astrophys (CAASTRO), Sydney, AUSTRALIA, pp. 85-88.
Maggs, CA & Musial-Gabrys, K 2017, 'LINNEAN SYSTEMATICS IN THE AGE OF BIG DATA', PHYCOLOGIA, INT PHYCOLOGICAL SOC, pp. 123-124.
Kajdanowicz, T, Michalski, R, Musial, K & Kazienko, P 2016, 'Learning in unlabeled networks - An active learning and inference approach', AI COMMUNICATIONS, IOS PRESS, pp. 123-148.View/Download from: Publisher's site
Abdullaev, S, McBurney, P & Musial, K 2016, 'Pricing Options with Portfolio-Holding Trading Agents in Direct Double Auction', ECAI 2016: 22ND EUROPEAN CONFERENCE ON ARTIFICIAL INTELLIGENCE, 22nd European Conference on Artificial Intelligence (ECAI), IOS PRESS, Hague, NETHERLANDS, pp. 1754-1755.View/Download from: Publisher's site
Gao, F & Musial-Gabrys, K 2016, 'Hybrid Structure-based Link Prediction Model', PROCEEDINGS OF THE 2016 IEEE/ACM INTERNATIONAL CONFERENCE ON ADVANCES IN SOCIAL NETWORKS ANALYSIS AND MINING ASONAM 2016, 8th IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM), IEEE, San Francisco, CA, pp. 1221-1228.
Venkata, SK, Keppens, J & Musial, K 2016, 'Agent based simulation to evaluate adaptive caching in distributed databases', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 455-462.View/Download from: Publisher's site
© Springer International Publishing Switzerland 2016. Caching frequently used data is a common practice to improve query performance in database systems. But traditional algorithms used for cache management prove to be insufficient in distributed environment where groups of users require similar or related data from multiple databases. Repeated data transfers can become a bottleneck leading to long query response time and high resource utilization. Our work focuses on adaptive algorithms to decide on optimal grain of data to be cached and cache refreshment techniques to reduce data transfers. In this paper, we present agent based simulation to investigate and in consequence improve cache management in the distributed database environment. Dynamic grain size and decisions on cache refreshment are made as a result of coordination and interaction between agents. Initial results show better response time and higher data availability compared to traditional caching techniques.
Guo, M, Yang, K, Musial-Gabrys, K, Min, G, Yin, H, Nguyen, NP, Jiang, Y, Kourtellis, N, Cheng, X, Leng, S, Wang, H & Dokoohaki, N 2015, 'Message from the MSNCom 2015 workshop chairs', Proceedings - 15th IEEE International Conference on Computer and Information Technology, CIT 2015, 14th IEEE International Conference on Ubiquitous Computing and Communications, IUCC 2015, 13th IEEE International Conference on Dependable, Autonomic and Secure Computing, DASC 2015 and 13th IEEE International Conference on Pervasive Intelligence and Computing, PICom 2015, p. lvi.View/Download from: Publisher's site
Abdullaev, S, Mcburney, P & Musial, K 2015, 'Direct exchange mechanisms for option pricing', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 269-284.View/Download from: Publisher's site
© Springer International Publishing Switzerland 2015. This paper presents the design and simulation of direct exchange mechanisms for pricing European options. It extends McAfee's single-unit double auction to multi-unit format, and then applies it for pricing options through aggregating agent predictions of future asset prices. We will also propose the design of a combinatorial exchange for the simulation of agents using option trading strategies. We present several option trading strategies that are commonly used in real option markets to minimise the risk of future loss, and assume that agents can submit them as a combinatorial bid to the market maker. We provide simulation results for proposed mechanisms, and compare them with existing Black-Scholes model mostly used for option pricing. The simulation also tests the effect of supply and demand changes on option prices. It also takes into account agents with different implied volatility. We also observe how option prices are affected by the agents' choices of option trading strategies.
Bródka, P, Magnani, M & Musial, K 2014, 'Message from SNAA 2014 program chairs', ASONAM 2014 - Proceedings of the 2014 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, p. xxxiv.View/Download from: Publisher's site
Krol, D, Budka, M & Musial, K 2014, 'Simulating the information diffusion process in complex networks using push and pull strategies', 2014 EUROPEAN NETWORK INTELLIGENCE CONFERENCE (ENIC), 1st European Network Intelligence Conference (ENIC), IEEE, Wroclaw, POLAND, pp. 1-8.View/Download from: Publisher's site
Kajdanowicz, T, Michalski, R, Musial, K & Kazienko, P 2013, 'Active Learning and Inference Method for Within Network Classification', 2013 IEEE/ACM INTERNATIONAL CONFERENCE ON ADVANCES IN SOCIAL NETWORKS ANALYSIS AND MINING (ASONAM), IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM), IEEE, Niagara Falls, CANADA, pp. 1299-1306.
Musial, K, Gabrys, B & Buczko, M 2013, 'What kind of network are you? - Using local and global characteristics in network categorisation tasks', 2013 IEEE/ACM INTERNATIONAL CONFERENCE ON ADVANCES IN SOCIAL NETWORKS ANALYSIS AND MINING (ASONAM), IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM), IEEE, Niagara Falls, CANADA, pp. 1366-1373.
Juszczyszyn, K, Gonczarek, A, Tomczak, JM, Musial, K & Budka, M 2012, 'A Probabilistic Approach to Structural Change Prediction in Evolving Social Networks', 2012 IEEE/ACM INTERNATIONAL CONFERENCE ON ADVANCES IN SOCIAL NETWORKS ANALYSIS AND MINING (ASONAM), IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, IEEE, Kadir Has Univ, Istanbul, TURKEY, pp. 996-1001.View/Download from: Publisher's site
Budka, M, Musial, K & Juszczyszyn, K 2012, 'Predicting the evolution of social networks: Optimal time window size for increased accuracy', Proceedings - 2012 ASE/IEEE International Conference on Privacy, Security, Risk and Trust and 2012 ASE/IEEE International Conference on Social Computing, SocialCom/PASSAT 2012, pp. 21-30.View/Download from: Publisher's site
This study investigates the data preparation process for predictive modelling of the evolution of complex networked systems, using an e - mail based social network as an example. In particular, we focus on the selection of optimal time window size for building a time series of network snapshots, which forms the input of chosen predictive models. We formulate this issue as a constrained multi - objective optimization problem, where the constraints are specific to a particular application and predictive algorithm used. The optimization process is guided by the proposed Windows Incoherence Measures, defined as averaged Jensen-Shannon divergences between distributions of a range of network characteristics for the individual time windows and the network covering the whole considered period of time. The experiments demonstrate that the informed choice of window size according to the proposed approach allows to boost the prediction accuracy of all examined prediction algorithms, and can also be used for optimally defining the prediction problems if some flexibility in their definition is allowed. © 2012 IEEE.
Musial, K & Sastry, N 2012, 'Social media - Are they underpinned by social or interest-based interactions?', ACM International Conference Proceeding Series, pp. 1-6.View/Download from: Publisher's site
On many social media and user-generated content sites, users can not only upload content but also create links with other users to follow their activities. It is interesting to ask whether the resulting user-user Followers' Network is based more on social ties, or shared interests in similar content. This paper reports our preliminary progress in answering this question using around five years of data from social video-sharing site vimeo. Many links in the Followers' Network are between users who do not have any videos in common, which would imply the network is not interest-based, but rather has a social character. However, the Followers' Network also exhibits properties unlike other social networks, for instance, clustering co-efficient is low, links are frequently not reciprocated, and users form links across vast geographical distances. In addition, analysis of the relationship strength, calculated as the number of commonly liked videos, people who follow each other and share some "likes" have more video likes in common than the general population. We conclude by speculating on the reasons for these differences and proposals for further work. © 2012 ACM.
Kazienko, P, Musial, K, Kukla, E, Kajdanowicz, T & Brodka, P 2011, 'Multidimensional Social Network: Model and Analysis', COMPUTATIONAL COLLECTIVE INTELLIGENCE: TECHNOLOGIES AND APPLICATIONS, PT I, 3rd International Conference on Computational Collective Intelligence (ICCCI 2011), SPRINGER-VERLAG BERLIN, Gdynia Maritime Univ, Gdynia, POLAND, pp. 378-387.
Kazienko, P, Kukla, E, Musial, K, Kajdanowicz, T, Brodka, P & Gaworecki, J 2011, 'A Generic Model for a Multidimensional Temporal Social Network', E-TECHNOLOGIES AND NETWORKS FOR DEVELOPMENT, 1st International Conference on e-Technologies and Networks for Development (ICeND 2011), SPRINGER-VERLAG BERLIN, Inst Finance Management, Dar es Salaam, TANZANIA, pp. 1-14.
Bródka, P, Skibicki, K, Kazienko, P & Musiał, K 2011, 'A degree centrality in multi-layered social network', Proceedings of the 2011 International Conference on Computational Aspects of Social Networks, CASoN'11, pp. 237-242.View/Download from: Publisher's site
Multi-layered social networks reflect complex relationships existing in modern interconnected IT systems. In such a network each pair of nodes may be linked by many edges that correspond to different communication or collaboration user activities. Multi-layered degree centrality for multi-layered social networks is presented in the paper. Experimental studies were carried out on data collected from the real Web 2.0 site. The multi-layered social network extracted from this data consists of ten distinct layers and the network analysis was performed for different degree centralities measures. © 2011 IEEE.
Juszczyszyn, K, Musiał, K & Budka, M 2011, 'Link prediction based on Subgraph evolution in dynamic social networks', Proceedings - 2011 IEEE International Conference on Privacy, Security, Risk and Trust and IEEE International Conference on Social Computing, PASSAT/SocialCom 2011, pp. 27-34.View/Download from: Publisher's site
We propose a new method for characterizing the dynamics of complex networks with its application to the link prediction problem. Our approach is based on the discovery of network subgraphs (in this study: triads of nodes) and measuring their transitions during network evolution. We define the Triad Transition Matrix (TTM) containing the probabilities of transitions between triads found in the network, then we show how it can help to discover and quantify the dynamic patterns of network evolution. We also propose the application of TTM to link prediction with an algorithm (called TTM-predictor) which shows good performance, especially for sparse networks analyzed in short time scales. The future applications and research directions of our approach are also proposed and discussed. © 2011 IEEE.
Juszczyszyn, K, Budka, M & Musial, K 2011, 'The Dynamic Structural Patterns of Social Networks Based on Triad Transitions', 2011 INTERNATIONAL CONFERENCE ON ADVANCES IN SOCIAL NETWORKS ANALYSIS AND MINING (ASONAM 2011), International Conference on Advances in Social Networks Analysis and Mining (ASONAM), IEEE COMPUTER SOC, Kaohsiung, TAIWAN, pp. 581-586.View/Download from: Publisher's site
Kazienko, P, Bródka, P, Musial, K & Gaworecki, J 2010, 'Multi-layered social network creation based on bibliographic data', Proceedings - SocialCom 2010: 2nd IEEE International Conference on Social Computing, PASSAT 2010: 2nd IEEE International Conference on Privacy, Security, Risk and Trust, pp. 407-412.View/Download from: Publisher's site
A method for extraction of the multi-layered social network based on the data about human collaborative achievements, in particular scientific papers, is presented in the paper. The objects linking people form a hierarchy, which is flattened in the pre-processing stage. Only one level of the hierarchy remains together with new activities moved from its other levels. Separate layers of the multi-layered social network are created based on these pre-processed activities. © 2010 IEEE.
Kazienko, P, Brodka, P & Musial, K 2010, 'Individual neighbourhood exploration in complex multi-layered social network', Proceedings - 2010 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT 2010, pp. 5-8.View/Download from: Publisher's site
Social networks can be extracted from different data about communication or common activities in organizations, companies or various Internet-based services. Different types of data processed may result in creation of separate layers in the complex multi-layered social network. Analysis of neighbourhoods of network members and their utilization to social group discovery appears to be an interesting and important research domain. Since there is no measure to evaluate structure of the neighbourhoods in the multi-layered social network, a new measure called cross layered multi-layered clustering coefficient (CLMCC) is proposed in the paper. It enables to analyse the density of mutual connections of neighbours that occur in at least a given number of layers in a social network. Additionally, experimental studies on real-world data are presented. © 2010 IEEE.
Bródka, P, Musial, K & Kazienko, P 2010, 'A method for group extraction in complex social networks', Communications in Computer and Information Science, pp. 238-247.View/Download from: Publisher's site
The extraction of social groups from social networks existing among employees in the company, its customers or users of various computer systems became one of the research areas of growing importance. Once we have discovered the groups, we can utilise them, in different kinds of recommender systems or in the analysis of the team structure and communication within a given population. The shortcomings of the existing methods for community discovery and lack of their applicability in multi-layered social networks were the inspiration to create a new group extraction method in complex multi-layered social networks. The main idea that stands behind this new concept is to utilise the modified version of a measure called by authors multi-layered clustering coefficient. © 2010 Springer-Verlag.
Juszczyszyn, K, Musiał, A, Musiał, K & Bródka, P 2010, 'Utilizing dynamic molecular modelling technique for predicting changes in complex social networks', Proceedings - 2010 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT 2010, pp. 1-4.View/Download from: Publisher's site
We present a method that utilises dynamic molecular modelling technique to track the changes within complex social network. The users forming a social network are interpreted as large sets of interacting particles. The data for the conducted research was obtained from e-mail communication within Enron company. The social network of employees was extracted and used to evaluate the methodology of social network dynamics modelling. © 2010 IEEE.
Brodka, P, Musial, K & Kazienko, P 2009, 'A Performance of Centrality Calculation in Social Networks', 2009 INTERNATIONAL CONFERENCE ON COMPUTATIONAL ASPECTS OF SOCIAL NETWORKS, PROCEEDINGS, International Conference on Computational Aspects of Social Networks, IEEE COMPUTER SOC, ESIGETEL, Fontainebleau, FRANCE, pp. 24-31.View/Download from: Publisher's site
Juszczyszyn, K, Musial, A, Musial, K & Brodka, P 2009, 'Molecular Dynamics Modelling of the Temporal Changes in Complex Networks', 2009 IEEE CONGRESS ON EVOLUTIONARY COMPUTATION, VOLS 1-5, IEEE Congress on Evolutionary Computation, IEEE, Trondheim, NORWAY, pp. 553-+.View/Download from: Publisher's site
Brodka, P, Musial, K & Kazienko, P 2009, 'Efficiency of Node Position Calculation in Social Networks', KNOWLEDGE-BASED AND INTELLIGENT INFORMATION AND ENGINEERING SYSTEMS, PT II, PROCEEDINGS, 13th International Conference on Knowledge-Based Intelligent Information and Engineering Systems, SPRINGER-VERLAG BERLIN, Univ Chile, Fac Phys Sci & Math, Santiago, CHILE, pp. 455-463.
Musiał, K, Kazienko, P & Bródka, P 2009, 'User position measures in social networks', Proceedings of the 3rd Workshop on Social Network Mining and Analysis, SNA-KDD '09.View/Download from: Publisher's site
Network analysis offers many centrality measures that are successfully utilized in the process of investigating the social network profile. The most important and representative measures are presented in the paper. It includes indegree centrality, proximity prestige, rank prestige, node position, outdegree centrality, eccentrality, closeness centrality, and betweenes centrality. Both feature analysis and experimental comparative studies revealed the general profile of selected measures.
Juszczyszyn, K & Musial, K 2009, 'Structural Changes in an Email-Based Social Network', AGENT AND MULTI-AGENT SYSTEMS: TECHNOLOGIES AND APPLICATIONS, PROCEEDINGS, 3rd KES International Symposium on Agent and Multi-Agent Systems, SPRINGER-VERLAG BERLIN, Uppsala Univ, Uppsala, SWEDEN, pp. 40-49.
Kazienko, P, Musial, K & Zgrzywa, A 2007, 'Evaluation of node position based on email communication', CONTROL AND CYBERNETICS, Conference on Data Processing Technologies, POLISH ACAD SCIENCES SYSTEMS RESEARCH INST, Poznan, POLAND, pp. 67-86.
Musial, K & Juszczyszyn, K 2009, 'Motif-based Analysis of Social Position Influence on Interconnection Patterns in Complex Social Network', 2009 FIRST ASIAN CONFERENCE ON INTELLIGENT INFORMATION AND DATABASE SYSTEMS, 1st Asian Conference on Intelligent Information and Database Systems, IEEE, Dong Hoi, VIETNAM, pp. 34-39.View/Download from: Publisher's site
Musial, K & Juszczyszyn, K 2009, 'Properties of Bridge Nodes in Social Networks', COMPUTATIONAL COLLECTIVE INTELLIGENCE: SEMANTIC WEB, SOCIAL NETWORKS AND MULTIAGENT SYSTEMS, 1st International Conference on Computational Collective Intelligence, SPRINGER-VERLAG BERLIN, Wroclaw, POLAND, pp. 357-364.
Musial, K, Juszczyszyn, K, Gabrys, B & Kazienko, P 2008, 'Patterns of Interactions in Complex Social Networks Based on Coloured Motifs Analysis', ADVANCES IN NEURO-INFORMATION PROCESSING, PT II, 15th International Conference on Neuro-Information Processing, SPRINGER-VERLAG BERLIN, Auckland, NEW ZEALAND, pp. 607-+.
Kazienko, P, Musiał, K & Kajdanowicz, T 2008, 'Profile of the social network in photo sharing systems', 14th Americas Conference on Information Systems, AMCIS 2008, pp. 2815-2826.
People, who interact, cooperate or share common activities within the photo sharing system can be seen as a multirelational social network. The results of their activities, i.e. tags, comments, references to favourites and others that semantically connect users through multimedia objects, i.e. pictures are the crucial component of the semantic web concept. Every online sharing system provides data that can be used for extraction of different kinds of relations grouped in layers in the multirelational social network. Layers and their profiles were identified and studied on two, spanned in time, snapshots of Flickr population for better understanding of social network structure complexity. Additionally, for each of the identified layers, a separate strength measure was proposed in the paper. The experiments on the Flickr photo sharing system revealed that users are inspired by both the semantic relationships between objects they operate on and social links they have to other users. Moreover, the density and affluence of the social network grows over course of time.
Musial, K, Kazienko, P & Kajdanowicz, T 2008, 'Social recommendations within the multimedia sharing systems', EMERGING TECHNOLOGIES AND INFORMATION SYSTEMS FOR THE KNOWLEDGE SOCIETY, PROCEEDINGS, 1st World Summit on the Knowledge Society (WSKS 2008), SPRINGER-VERLAG BERLIN, Athens, GREECE, pp. 364-+.
Kazienko, P & Musial, K 2008, 'Mining personal social features in the community of email users', SOFSEM 2008: THEORY AND PRACTICE OF COMPUTER SCIENCE, 34th Conference on Current Trends in Theory and Practice of Computer Science, SPRINGER-VERLAG BERLIN, Novy Smokovec, SLOVAKIA, pp. 708-719.
Kazienko, P, Musial, K & Juszczyszyn, K 2008, 'Recommendation of multimedia objects based on similarity of ontologies', KNOWLEDGE - BASED INTELLIGENT INFORMATION AND ENGINEERING SYSTEMS, PT 1, PROCEEDINGS, 12th International Conference on Knowledge-Based Intelligent Information and Engineering Systems, SPRINGER-VERLAG BERLIN, Zagreb, CROATIA, pp. 194-201.
Juszczyszyn, K, Kazienko, P & Musial, K 2008, 'Local topology of social network based on motif analysis', KNOWLEDGE-BASED INTELLIGENT INFORMATION AND ENGINEERING SYSTEMS, PT 2, PROCEEDINGS, 12th International Conference on Knowledge-Based Intelligent Information and Engineering Systems, SPRINGER-VERLAG BERLIN, Zagreb, CROATIA, pp. 97-105.
Kazienko, P & Musial, K 2007, 'Assessment of personal importance based on social networks', MICAI 2007: ADVANCES IN ARTIFICIAL INTELLIGENCE, 6th Mexican International Conference on Artificial Intelligence (MICAI 2007), SPRINGER-VERLAG BERLIN, Aguascalientes, MEXICO, pp. 529-+.
Kazienko, P & Musial, K 2006, 'Social capital in Online social networks', KNOWLEDGE-BASED INTELLIGENT INFORMATION AND ENGINEERING SYSTEMS, PT 2, PROCEEDINGS, 10th International Conference on Knowledge-Based and Intelligent Information and Engineering Systems, SPRINGER-VERLAG BERLIN, Bournemouth, ENGLAND, pp. 417-424.
Musial-Gabrys, K, Wahid-Ul-Ashraf, A & Budka, M 2019, 'Simulation and Augmentation of Social Networks for Building Deep Learning Models'.
Kajdanowicz, T, Michalski, R, Musiał, K & Kazienko, P, 'Learning in Unlabeled Networks - An Active Learning and Inference Approach'.
The task of determining labels of all network nodes based on the knowledge
about network structure and labels of some training subset of nodes is called
the within-network classification. It may happen that none of the labels of the
nodes is known and additionally there is no information about number of classes
to which nodes can be assigned. In such a case a subset of nodes has to be
selected for initial label acquisition. The question that arises is: "labels of
which nodes should be collected and used for learning in order to provide the
best classification accuracy for the whole network?". Active learning and
inference is a practical framework to study this problem.
A set of methods for active learning and inference for within network
classification is proposed and validated. The utility score calculation for
each node based on network structure is the first step in the process. The
scores enable to rank the nodes. Based on the ranking, a set of nodes, for
which the labels are acquired, is selected (e.g. by taking top or bottom N from
the ranking). The new measure-neighbour methods proposed in the paper suggest
not obtaining labels of nodes from the ranking but rather acquiring labels of
their neighbours. The paper examines 29 distinct formulations of utility score
and selection methods reporting their impact on the results of two collective
classification algorithms: Iterative Classification Algorithm and Loopy Belief
We advocate that the accuracy of presented methods depends on the structural
properties of the examined network. We claim that measure-neighbour methods
will work better than the regular methods for networks with higher clustering
coefficient and worse than regular methods for networks with low clustering
coefficient. According to our hypothesis, based on clustering coefficient we
are able to recommend appropriate active learning and inference method.
My activities resulted in maintaining contacts with various research centres including the commercial companies’ research divisions of BT, Research and Engineering Centre, SAS Institute, Badoo, Phorm, and Affectv. The collaboration with Affectv resulted in joint Open Graph Initiative where we organised open challenges based on large social network data sets that come from 50 different social data providers all over the Europe. During my work at KCL I also collaborated and signed a substantial consultancy contract with Badoo, a company which owns a dating-focused social discovery website and has over 170 million active users in 180 countries. Another consultancy contract was signed with Phorm company and its goal was to explore machine learning approaches to analyse company data.
01/2014 – 03/2014 - Consultancy contract with Phorm. The main goal of the contract is to help company to understand in greater detail the mechanics of company's information system. The title of the project is: “Machine Learning approaches to analyse company data”. Income: £16,000.
01/2013 – 03/2014 - Consultancy agreement with Badoo Trading Limited aiming at understanding customer related data. The main goals of the contract were (i) to understand in greater detail the mechanisms of their email-based system, (ii) to develop a meaningful measure of user churn, and (iii) to discover the intentions of users. Income: £20,000.
Badoo was also involved in my EPSRC First Grant Application and was prepared to financially support the project (£60,000).
06/2013 – 01/2016 - Collaboration with Affectv company resulted in Open Graph Initiative in which the company opens up its aggregated social data for external world.
My collaboration with Affectv also resulted in their involvement in the application for the Centre for Doctoral Training in Data Science led by Bournemouth University (Prof. Bogdan Gabrys) in 2013.