Prof. Hiroshi Iskikawa

Tokyo Metropolitan University

Japan

 (Short Biography)

Social big data mining: concepts and use cases

Right at this moment, data deluge is continuously producing a large amount of data in various sectors of our modern society. Such data contain data originating both in our physical real world and in social media and are collectively called social big data. If both kinds of data are analyzed in a mutually related fashion, especially with respect to space and time, values which cannot be acquired only by independent analysis will be discovered and utilized in various applications ranging from business to science. In this talk, an integrated framework for modeling and analyzing interactions involving both the physical real world and social media will be introduced together with a few applications as use cases.

 (Short Biography)

 
How to Use Large Computers to Process
Really Big Sequences and Graphs

Modern applications, including bioinformatics, stock market prediction, or intrusion detection in web servers, depend on the analysis of very large datasets in the form of long sequences (e.g., DNA) or large graphs (e.g., interactions among proteins). Because of the size and complexity of such datasets, processing requires very large computing infrastructures. This talk will give an overview of the work done in the InfoCloud group at KAUST. Our research focuses on the efficient indexing and implementation of generic operators, such as finding all frequent subgraphs in a large graph, that can be used by applications to speedup the data mining and knowledge extraction processes. We will show that our methods scale to orders of magnitude larger datasets than existing approaches, and can run on a variety of computing infrastructures, including stand-alone workstations, Linux clusters, cloud (e.g., Amazon EC2) and supercomputers with tens of thousands of processors.

Prof. Ernesto Damiani
EBTIC - Khalifa University - Abu Dhabi

UAE

 (Short Biography)

Big Data: Threat Landscape and Protection Gap Analysis

The advent of IoT and cloud services has resulted in collecting and sharing massive amounts of data. From a security perspective, these data represents a valuable target for attackers. As data-driven processes become integrated in the fabric of business, the entire society is becoming increasingly vulnerable to threats to data reliability and availability. Finally, the increase in redundancy of data available for collection, analysis and dissemination have strained traditional rules to protect privacy and confidentiality.
The Big Data threat landscape continues to evolve. Opportunistic one-shot attacks have been supplemented by leakages that are more persistent and, in many cases, far more worrisome. This means that we need to start designing Big Data systems not just to prevent attacks and recover from them, but also to detect successful attackers quickly and contain them so that any data leakage can be identified and countered. This talk starts by introducing the emerging Big Data Threat Landscape with reference to some vertical domains and performs a Protection Gap Analysis to list some known vulnerabilities. Then a paradigm of Detect, Contain and Recover is introduced as a practical foundation for managing risks connected to Big Data.