Tomasz Kajdanowicz

Balancing Speed and Coverage by Sequential Seeding in Complex Networks

Information spreading in complex networks is often modeled as diffusing information with certain probability from nodes that possess it to their neighbors that do not. Information cascades are triggered when the activation of a set of initial nodes – seeds – results in diffusion to large number of nodes. Here, several novel approaches for seed initiation that replace the commonly used activation of all seeds at once with a sequence of initiation stages are introduced. Sequential strategies at later stages avoid seeding highly ranked nodes that are already activated by diffusion active between stages. The gain arises when a saved seed is allocated to a node difficult to reach via diffusion. Sequential seeding and a single stage approach are compared using various seed ranking methods and diffusion parameters on real complex networks. The experimental results indicate that, regardless of the seed ranking method used, sequential seeding strategies deliver better coverage than single stage seeding in about 90% of cases. Longer seeding sequences tend to activate more nodes but they also extend the duration of diffusion. Various variants of sequential seeding resolve the trade-off between the coverage and speed of diffusion differently.

Using Graph and Vertex Entropy to Compare Empirical Graphs with Theoretical Graph Models

Over the years, several theoretical graph generation models have been proposed. Among the most prominent are: the Erdős–Renyi random graph model, Watts–Strogatz small world model, Albert–Barabasi preferential attachment model, Price citation model, and many more. Often, researchers working with real-world data are interested in understanding the generative phenomena underlying their empirical graphs. They want to know which of the theoretical graph generation models would most probably generate a particular empirical graph. In other words, they expect some similarity assessment between the empirical graph and graphs artificially created from theoretical graph generation models. Usually, in order to assess the similarity of two graphs, centrality measure distributions are compared. For a theoretical graph model this means comparing the empirical graph to a single realization of a theoretical graph model, where the realization is generated from the given model using an arbitrary set of parameters. The similarity between centrality measure distributions can be measured using standard statistical tests, e.g., the Kolmogorov–Smirnov test of distances between cumulative distributions. However, this approach is both error-prone and leads to incorrect conclusions, as we show in our experiments. Therefore, we propose a new method for graph comparison and type classification by comparing the entropies of centrality measure distributions (degree centrality, betweenness centrality, closeness centrality). We demonstrate that our approach can help assign the empirical graph to the most similar theoretical model using a simple unsupervised learning method.

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. nA 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 Propagation. nWe 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.

Priority rank model for social network generation

Currently available artificial network generation models are characterized by consistency and low variance due to the rigidity of models underlying assumptions. Networks generated from these models are usually too regular and do not contain noise and imbalance inherent in networks induced by human behavior. An important consequence is that much research on social network analysis presented in recent years used idealistic artificial networks that did not conform to reality. In order to alleviate this problem we introduce a new network generation model capable of modeling a broad spectrum of networks. In our model, the network formation process is not hard-coded into the model. Rather, we propose a simple mechanism for network creation based on priority ranking, and we encode the guiding principle of network formation as a distance function. By only changing the distance function definition and using the same priority ranking mechanism we are able to model very diverse networks. Our preliminary results show that we can mimic the behavior of popular artificial network generation models, such as the Erdös-Rényi random network model, the Watts-Strogatz small world model, or the Albert-Barabási preferential attachment model, but we can generate new types of networks as well. Following the principles of Open Science we publish the source code used to perform experiments and publish results in a public repository.

On Quality Assesement in Wikipedia Articles Based on Markov Random Fields

This article investigates the possibility of accurate quality prediction of resources generated by communities based on the crowd-generated content. We use data from Wikipedia, the prime example of community-run site, as our object of study. We define the quality as a distribution of user-assigned grades across a predefined range of possible scores and present a measure of distribution similarity to quantify the accuracy of a prediction. The proposed method of quality prediction is based on Markov Random Field and its Loopy Belief Propagation implementation. Based on our results, we highlight key problems in the approach as presented, as well as trade-offs caused by relying solely on network structure and characteristics, excluding metadata. The overall results of content quality prediction are promising in homophilic networks.

WordNet2Vec: Corpora agnostic word vectorization method

A complex nature of big data resources demands new methods for structuring especially for textual content. WordNet is a good knowledge source for comprehensive abstraction of natural language as its good implementations exist for many languages. Since WordNet embeds natural language in the form of a complex network, a transformation mechanism WordNet2Vec is proposed in the paper. It creates vectors for each word from WordNet. These vectors encapsulate general position - role of a given word towards all other words in the natural language. Any list or set of such vectors contains knowledge about the context of its component within the whole language. Such word representation can be easily applied to many analytic tasks like classification or clustering. The usefulness of the WordNet2Vec method was demonstrated in sentiment analysis, i.e. classification with transfer learning for the real Amazon opinion textual dataset.

Requirements and validation of a prototype learning health system for clinical diagnosis

Diagnostic error is a major threat to patient safety in the context of family practice. The patient safety implications are severe for both patient and clinician. Traditional approaches to diagnostic decision support have lacked broad acceptance for a number of well‐documented reasons: poor integration with electronic health records and clinician workflow, static evidence that lacks transparency and trust, and use of proprietary technical standards hindering wider interoperability. The learning health system (LHS) provides a suitable infrastructure for development of a new breed of learning decision support tools. These tools exploit the potential for appropriate use of the growing volumes of aggregated sources of electronic health records.

Incremental Learning in Dynamic Networks for Node Classification.

An incremental learning method for nodes’ classification is presented in the paper. In particular, there is proposed an active scheme algorithm for multi-class classification of nodes’ states that varies over time and depends on information spread in the network. Demonstration of the method is conducted using social network dataset. According to sent messages between nodes, the emotional state of the message sender updates each receiving node’s feature vector and the method tries to classify next emotional state of the receiver. The novelty of the proposed approach lies in applying incremental learning method for non-stationary network environment. There are demonstrated some properties of the proposed method in experiments with real data set, showing that the method can effectively classify the future state of nodes.