Retrieval of images from large image archives based solely on their visual similarity to a query image provides an exciting alternative to conventional text-based search. For content-based retrieval images are represented in terms of visual features. The question of how to combine these for similarity computation is typically addressed by eliciting relevance feedback from the user on the retrieved images. We argue in this thesis that the prevailing approach to relevance feedback suffers from three significant shortcomings: firstly, it leaves unsolved the question of how to combine features for the first retrieval; secondly, the advantage of automated content-extraction over manual annotation is greatest for large collections but if the query image is not constrained to come from the indexed collection, content-based retrieval entails imagewise comparisons leading to prohibitive response times; thirdly, users may only have vaguely defined information needs or may change their needs in the course of the interaction. The large majority of relevance feedback techniques are ill-suited for such undirected exploration. We propose a new framework of user interaction that addresses these limitations. It is centred on what we call the NN k idea. The NN k of an image are all those images that are most similar to it under some combination of features. They can be viewed as representatives of the possible

This paper proposes novel methods for visualizing specifically the large power-law graphs that arise in sociology and the sciences. In such cases a large portion of edges can be shown to be less important and removed while preserving component connectedness and other features (e.g. cliques) to more clearly reveal the network’s underlying connection pathways. This simplification approach deterministically filters (instead of clustering) the graph to retain important node and edge semantics, and works both automatically and interactively. The improved graph filtering and layout is combined with a novel computer graphics anisotropic shading of the dense crisscrossing array of edges to yield a full social network and scale-free graph visualization system. Both quantitative analysis and visual results demonstrate the effectiveness of this approach.

OLAP (On-Line Analytical Processing) is an important notion in data analysis. Recently, more and more graph or networked data sources come into being. There exists a similar need to deploy graph analysis from different perspectives and with multiple granularities. However, traditional OLAP technology cannot handle such demands because it does not consider the links among individual data tuples. In this paper, we develop a novel graph OLAP framework, which presents a multi-dimensional and multi-level view over graphs. The contributions of this work are two-fold. First, starting from basic definitions, i.e., what are dimensions and measures in the graph OLAP scenario, we develop a conceptual framework for data cubes on graphs. We also look into different semantics of OLAP operations, and classify the framework into two major subcases: informational OLAP and topological OLAP. Then, with more emphasis on informational OLAP (topological OLAP will be covered in a future study due to the lack of space), we show how a graph cube can be materialized by calculating a special kind of measure called aggregated graph and how to implement it efficiently. This includes both full materialization and partial materialization where constraints are enforced to obtain an iceberg cube. We can see that the aggregated graphs, which depend on the graph properties of underlying networks, are much harder to compute than their traditional OLAP counterparts, due to the increased structural complexity of data. Empirical studies show insightful results on real datasets and demonstrate the efficiency of our proposed optimizations.

Learning communities from a graph is an important problem in many domains. Different types of communities can be generalized as link-pattern based communities. In this paper, we propose a general model based on graph approximation to learn link-pattern based community structures from a graph. The model generalizes the traditional graph partitioning approaches and is applicable to learning various community structures. Under this model, we derive a family of algorithms which are flexible to learn various community structures and easy to incorporate the prior knowledge of the community structures. Experimental evaluation and theoretical analysis show the effectiveness and great potential of the proposed model and algorithms. 1

Abstract. In many areas of life science, such as biology and medicine, ontologies are nowadays commonly used to annotate objects of interest, such as biological samples, clinical pictures, or species in a standardized way. In these applications, an ontology is merely a structured vocabulary in the form of a tree or a directed acyclic graph of concepts. Typically, ontologies are stored together with the data they annotate in relational databases. Querying such annotations must obey the special semantics encoded in the structure of the ontology, i.e. relationships between terms, which is not possible using standard SQL alone. In this paper, we develop a new method for querying DAGs using a pre-computed index structure. Our new indexing method extends the pre- / postorder ranking scheme, which has been studied intensively for trees, to DAGs. Using typical queries on ontologies, we compare our approach to two other commonly used methods, i.e., a recursive database function and the pre-computation of the transitive closure of a DAG. We show that pre-computed indexes are an order of magnitude faster than recursive methods. Clearly, our new scheme is slower than usage of the transitive closure, but requires only a fraction of the space and is therefore applicable even for very large ontologies with more than 200,000 concepts. 1

Words in natural language documents exist as a small world network. Thus the extensive physics algorithms for extracting community structure are applicable. We present a novel method for semantically clustering a large collection of documents using small world communities. We combine modified physics algorithms with traditional information retrieval techniques. A term network is generated from the document collection, the terms are clustered into small world communities, the semantic term clusters are used to generate overlapping document clusters. The algorithm combines the speed of single link with the quality of complete link. Clustering takes place in nearly real-time and the results are judged to be coherent by expert users.

Proceedings A coherent graph-based semantic clustering and summarization approach for biomedical literature and a new summarization evaluation method

The hierarchical edge bundle method clusters the graph edges to better understand and analyze graphs, but its effectiveness relies critically on the quality of the hierarchical organization of its nodes and edges. This paper proposes a novel graph visualization approach that extracts the community structure of a network and organizes it into a more balanced and meaningful hierarchy so that its edge bundle rendering better indicates its structure. Results on several data sets demonstrate that this approach clarifies realworld communication, collaboration and competition network structure and reveals information missed in previous visualizations.

The Web has connected millions of users by various communication tools for social purposes. Daily, huge amount of social data are being created through fingertips, driven by various of social actions that involve a wide range of user-produced content (e.g. emails or collaborative documentations). Often being a part of many contemporary Web applications, user social networks are gaining increasing attention from both industry and academia as they seem to have become a promising vehicle for delivering better user experience. Accordingly, computational social network analysis has become an important topic in user data mining. Despite a long history of structural social network analysis and recent interests in user behavior analysis, little research has addressed social contents in social networks and heterogeneous networks in user behavior. In fact, social content and heterogeneity are two key elements in contemporary online social networks that offer great benefits: social contents provide more semantic information; meanwhile heterogeneous social networks allow diversified perception of users. Motivated by these considerations, this dissertation seeks to improve traditional computational