Keynote Speech I
Flexible and Secure Cloud Computing: The FlexCloud Approach (Tentative)
Cloud computing is a very common trend and enables users to dynamically access server resources on demand in order to safe infrastructure costs and to shift responsibilities from the user client to a backend infrastructure provider. However, it also comes with many challenges due to proprietary solutions, due to issues of trust and reliability, and due to lack of control regarding the information flow.
The talks outlines the potential of cloud computing and also its challenges. It then presents a novel solution name FlexCloud that acts as a smart intermediary between the end user and different cloud providers in the internet or intranet. The main goal is to implement the vision of a so-called p-cloud (or personal cloud) that can be customized according to the user requirements and the given environment. Most importantly, confidentiality, integrity, and availability of processed information are addressed. In addition to the selective use of cryptographic mechanisms, systematic information dispersal algorithms are employed in order to make cloud storage more reliable.
Eventually, it is shown how the approach can be generalized towards a full cloud lifecycle vision with service-level agreements, dynamic cloud configuration, cloud monitoring at runtime, and cloud adaptation and optimization on demand. The results are backed by a prototype implementation of a secure cloud storage system within the FlexCloud project.
Keynote Speech II
Link Mining and Information Retrieval from Complex Networks (Tentative)
Search engines such as Google and Yahoo have been widely and commonly used to retrieve information from World Wide Web. These search engines work well when one has concrete and exact keywords to identify the related pages. However, there is a case that one cannot successfully find the exact keywords to retrieve information which he or she wants to find. Also, there will be a case that one wants to find related pages that does not necessarily contain the original keywords that is used to invoke the search. In these cases one sometimes needs to perform so called net surfing, which is usually a time consuming task, to find related information. In this talk, we present alternative approach for information retrieval to cope with these cases.
The recent study of networks in computer science, such as computer networks, the WWW, and social network services, shows that very large real-world networks are networks with non-trivial topological features (and are sometimes called complex network). These features include a heavy tail in the degree distribution, a high clustering coefficient, and hierarchical structures. Among these features, we are especially interested in community structure of complex network. A network is said to have community structure if the nodes of the network can be grouped into sets of nodes such that each set of nodes is densely connected internally. In WWW, if we can find a community of pages that includes the page one is initially interested in, one might successfully be able to get access to other highly related pages that don't necessarily include the keywords.
Several efficient community detection algorithms have been proposed. However, these algorithms tend to detect a excessive community that includes pages that are less related. We have devised a couple of methods to improve precision and recall of community detection. These methods and some empirical results obtained from experiments based on the Wikipedia are presented.